From haas at xmission.com Wed Jul 1 08:28:46 2009 From: haas at xmission.com (Walt Haas) Date: Wed Jul 1 09:17:58 2009 Subject: [UPHPU] Using a Database for Object Storage (Sad Times) In-Reply-To: <013a01c9f9ec$92a00b10$b7e02130$@com> References: <013a01c9f9ec$92a00b10$b7e02130$@com> Message-ID: <4A4B729E.7040609@xmission.com> You can't really protect from concurrent updates at the PHP level. This problem is normally solved by using the MySQL two-phase commit mechanism: acquire all the data needed for the update, start a MySQL transaction, update the DB, then commit the update. See http://dev.mysql.com/doc/refman/5.0/en/sql-syntax-transactions.html -- Walt Kirk Ouimet wrote: > Hi List, > > > > I am currently storing an object in a MySQL database. The object is > continually being updated by many (possibly simultaneous) requests. Workflow > looks something like this: > > > > 1. Data comes in from one of many sources > > 2. PHP reads the object from the database > > 3. PHP adds data to the object > > 4. PHP writes the object back to the database > > > > I think I'm running into a concurrency issue where I have record of new data > being received, but the database object does not reflect the new data. I > feel like I need to make the object a mutex but I have no idea how to go > about doing that with PHP. Can anyone provide alternative designs that will > be a little more robust and handle concurrency? From lists at kittypee.com Wed Jul 1 09:40:45 2009 From: lists at kittypee.com (Lonnie Olson) Date: Wed Jul 1 10:29:57 2009 Subject: [UPHPU] Using a Database for Object Storage (Sad Times) In-Reply-To: <013a01c9f9ec$92a00b10$b7e02130$@com> References: <013a01c9f9ec$92a00b10$b7e02130$@com> Message-ID: <8bcade370907010840x18df024o778b444b6aaa4843@mail.gmail.com> On Tue, Jun 30, 2009 at 7:38 PM, Kirk Ouimet wrote: > I am currently storing an object in a MySQL database. The object is > continually being updated by many (possibly simultaneous) requests. Workflow > looks something like this: > > > > 1. ? ? ? Data comes in from one of many sources > > 2. ? ? ? PHP reads the object from the database > > 3. ? ? ? PHP adds data to the object > > 4. ? ? ? PHP writes the object back to the database How are you storing your object in the database? If you are storing the whole object in a single column via serialize or some other method, you will have this problem. This problem doesn't really exist if you are using "classic" database usage. You may just need to change how your object is stored. Each data point should have it's own column. Then your workflow will go something like this: 1. Data comes in 2. PHP reads current data 3. PHP adds data to object 4. PHP writes back only the data points that are changed Depending on what you do with the object, you may be able to skip steps 2 and 3. Now, the only concurrency problem is when two requests wish to modify the same data point, but that is a whole other problem. --lonnie From miles.togoe at gmail.com Wed Jul 1 09:58:21 2009 From: miles.togoe at gmail.com (MilesTogoe) Date: Wed Jul 1 10:48:45 2009 Subject: [UPHPU] Using a Database for Object Storage (Sad Times) In-Reply-To: <8bcade370907010840x18df024o778b444b6aaa4843@mail.gmail.com> References: <013a01c9f9ec$92a00b10$b7e02130$@com> <8bcade370907010840x18df024o778b444b6aaa4843@mail.gmail.com> Message-ID: <4A4B879D.30204@gmail.com> On 07/01/2009 09:40 AM, Lonnie Olson wrote: > On Tue, Jun 30, 2009 at 7:38 PM, Kirk Ouimet wrote: > >> I am currently storing an object in a MySQL database. The object is >> continually being updated by many (possibly simultaneous) requests. Workflow >> looks something like this: >> >> >> >> 1. Data comes in from one of many sources >> >> 2. PHP reads the object from the database >> >> 3. PHP adds data to the object >> >> 4. PHP writes the object back to the database >> > > How are you storing your object in the database? If you are storing > the whole object in a single column via serialize or some other > method, you will have this problem. > > This problem doesn't really exist if you are using "classic" database > usage. You may just need to change how your object is stored. Each > data point should have it's own column. > > Then your workflow will go something like this: > 1. Data comes in > 2. PHP reads current data > 3. PHP adds data to object > 4. PHP writes back only the data points that are changed > > Depending on what you do with the object, you may be able to skip steps 2 and 3. > Now, the only concurrency problem is when two requests wish to modify > the same data point, but that is a whole other problem. > > --lonnie > > > or for storing objects, you might consider a key:value database such as couchdb, mongodb, etc - depends on what exactly your transactions are like as to whether speed will compare with a relatonal db. From alvaro at epliant.com Wed Jul 1 10:38:49 2009 From: alvaro at epliant.com (Alvaro Carrasco) Date: Wed Jul 1 11:28:05 2009 Subject: [UPHPU] Using a Database for Object Storage (Sad Times) In-Reply-To: <013a01c9f9ec$92a00b10$b7e02130$@com> References: <013a01c9f9ec$92a00b10$b7e02130$@com> Message-ID: <4A4B9119.90500@epliant.com> Hi Kirk, Kirk Ouimet wrote: > Hi List, > > > > I am currently storing an object in a MySQL database. The object is > continually being updated by many (possibly simultaneous) requests. Workflow > looks something like this: > > > > 1. Data comes in from one of many sources > > 2. PHP reads the object from the database > > 3. PHP adds data to the object > > 4. PHP writes the object back to the database > > > > I think I'm running into a concurrency issue where I have record of new data > being received, but the database object does not reflect the new data. I > feel like I need to make the object a mutex but I have no idea how to go > about doing that with PHP. Can anyone provide alternative designs that will > be a little more robust and handle concurrency? > > Try wrapping the select and the update in one transaction. And doing a SELECT ... FOR UPDATE: $pdo->beginTransaction(); $stmt = $pdo->query("SELECT obj FROM table_with_object WHERE object_id = 1 FOR UPDATE"); // ... // get the object out of the statement and have php add the data to it // ... $pdo->exec("UPDATE table_with_object SET object = '$obj' WHERE object_id = 1"); $pdo->commit(); This is just a guess, i haven't tested it. Also make sure you use prepared statements to prevent SQL injection, I didn't here for the sake of brevity. Alvaro From bigdog at venticon.com Wed Jul 1 10:41:53 2009 From: bigdog at venticon.com (thebigdog) Date: Wed Jul 1 11:31:07 2009 Subject: [UPHPU] Using a Database for Object Storage (Sad Times) In-Reply-To: <013a01c9f9ec$92a00b10$b7e02130$@com> References: <013a01c9f9ec$92a00b10$b7e02130$@com> Message-ID: <4A4B91D1.6060900@venticon.com> > I am currently storing an object in a MySQL database. The object is > continually being updated by many (possibly simultaneous) requests. Workflow > looks something like this: > > 1. Data comes in from one of many sources > > 2. PHP reads the object from the database > > 3. PHP adds data to the object > > 4. PHP writes the object back to the database > > I think I'm running into a concurrency issue where I have record of new data > being received, but the database object does not reflect the new data. I > feel like I need to make the object a mutex but I have no idea how to go > about doing that with PHP. Can anyone provide alternative designs that will > be a little more robust and handle concurrency? This is actually a concurrency issue that spans the database and php. You will need to create some type of locking that will allow a request to lock the db for read/write as it completes it "unit of work" and in your php code. Most languages suffer from this type of scenario. This also introduces some performance issues as well. For example, if you lock on the db level then you will have requests that block as the workflow tries to complete for each request. As you are using mysql, then I would suggest that you use the locking in mysql and propagate that information up to our php objects. Here is a pretty easy example of what I am talking about: http://www.perplexedlabs.com/2008/02/06/mutex-with-php-and-mysql/ -- thebigdog From herlo1 at gmail.com Mon Jul 6 17:54:48 2009 From: herlo1 at gmail.com (Clint Savage) Date: Mon Jul 6 18:43:52 2009 Subject: [UPHPU] Utah Open Source Monthly News/Events In-Reply-To: References: Message-ID: The Utah Open Source Foundation, in its goal to promote Open Source throughout Utah and beyond, is proud to promote the following events. These events are generally related to Open Source and Technology in Utah. ?If your event is not listed below, please contact clint@utos.org to get it added. Utah Technology Community Announcements - Utah Open Source is looking for more volunteers to join our 'Core Team'. ?Position definitions and contacts are available at http://wiki.utos.org/All_Positions. - UTOSC 2009 'Call for Papers' is now open. ?Submit your abstract - http://2009.utosc.com - UTOS-ConMan HackNights are still hapening every Tuesday night in June, watch http://utos.org for announcements. - PodCampSLC 2010 Kick-off Meeting will be held July 13 @ 7pm. ?If you are interested in helping contact podcampslc@gmail.com If you have announcements or events for August 2009 or beyond happening around open source in the Intermountain West which you would like included, please contact clint@utos.org. ------------------------------------------------------------------------------------------------------- Upcoming Activities for July 2009 Wed, July 8, 11:30pm ? 1:00pm SLLUG: Daytime SIG Meeting Topic: Cobbler Where: BetaLoft SLC - 357 West 200 S, Suite 201, Salt Lake City Contact: Clint Savage, herlo1@gmail.com Wed, July 8, 7:30pm ? 9:00pm Provo Linux Users Group Topic: KVM & QEMU Link: http://plug.org Where: Omniture Contact: Ryan Simpkins, simpkins.ryan @gmail.com Thu, July 9, 6pm ? 9pm Utah Mobile Developers Group Where: STG Dev Center, 555 South 300 East, Salt Lake City, Utah Contact: Glen Lewis, glen @ glenlewis.com Thu, July 9, 7:00pm ? 8:30pm Utah Python User Group Meeting Link: http://utahpython.org Topic: GUI Toolkit Extravaganza cont'd Where: University of Utah, Emma Eccles Jones Medical Research Building, Room 1200 Contact: Dave, tonedevf AT gmail.com Sat, July 11th, 12 - 5pm Utah CodeAway Link: http://codeaway.org/ Where: TBD Contact: Laura Moncur, laura @ moncur.biz Sat, July 11, 6pm ? 8pm Ubuntu-Utah Meeting Link: http://utah.ubuntu-us.org Where: University of Utah - Merrill Engineering Building (MEB) Comp-Sci Labs Rm 2555 Contact: Aaron Toponce, aarontoponce@ gmail.com Wed, July 15, 7:10pm ? 8:40pm Salt Lake Linux User Group (SLLUG) Link: http://sllug.org Where: University of Utah, Warnock Engineering Building (WEB) room 101 or 103 (Previously known as EMCB) Contact: Marc Christensen marc aT mecworks.com Thu, July 16, 6pm ? 9pm Utah Java User's Group (ujug.org) Link: http://ujug.org Where: Intermountain Medical Center, Doty Family Education Center, Murray, Utah Contact: Chris Maki, chrismaki AT me.com Thu, July 16, 7:00pm ? 8:30pm Utah PHP User Group (UPHPU) Link: http://uphpu.org Where: Bill Good Marketing, Draper, Utah Contact: Victor Villa, vvilla @ gmail.com Thu, July 23, 6pm ? 8pm Logan Dev Group Where: Room 208 (Faculty Seminar Lounge), Merrill-Cazier Library, Utah State University, Logan, UT Contact: Matthew Reinbold, matthew.reinbold At voxpopdesign.com Tue, July 28, 6:30pm ? 8:00pm SLC Ruby (slc.rb) Link: http://groups-beta.google.com/group/urug Where: Neumont University Room #300 Contact: Jake Mallory, tinomen @gmail.com Tue, July 28, 7:30pm - 9:00pm Ogden Area Linux User Group Link: http://oalug.com Where: Weber County Main Library, Board Room Contact: Seth House, whiteinge@ gmail.com Fri, July 31, 12:30pm ? 2pm UTOS/UPHPU Geek Lunch Link: http://utos.org/geek-lunch Contact: Victor Villa, vvilla AT gmail.com ------------------------------------------------------------------------------------------------------- Each of the events are from the Utah Tech Events calendar. ?Feel free to subscribe by clicking the link below (or adding the following feed to your calendaring program) ------------------------------------------------------------------------------------------------------- Utah Tech Events Calendar - Google Calendar: http://sn.im/ute-calendar Utah Tech Events Calendar - iCal Feed: http://sn.im/ute-calendar-ics Utah Open Source Foundation also regularly records Local User Group (LUG) meetings throughout the state. ?Feel free to check out our podcasts and live streaming schedules. Utah Open Source Podcasts - http://podcast.utos.org From wadeshearer.lists at me.com Wed Jul 8 18:56:21 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Wed Jul 8 19:45:23 2009 Subject: [UPHPU] submitting a form with javascript, dynamic form name Message-ID: <076B1073-452F-4EBB-A17D-4982F16CA774@me.com> I am submitting a form with javascript like this? document.formname.submit(); I need the name of the form to be a variable though. Is that possible? Meaning, that "formname" should be a variable instead of the actual name of the form. I tried? document. + formname + .submit(); ?but that didn't work. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090708/cb6a14e1/smime.bin From wadeshearer.lists at me.com Wed Jul 8 18:59:57 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Wed Jul 8 19:48:59 2009 Subject: [UPHPU] submitting a form with javascript, dynamic form name In-Reply-To: <076B1073-452F-4EBB-A17D-4982F16CA774@me.com> References: <076B1073-452F-4EBB-A17D-4982F16CA774@me.com> Message-ID: <2E568B47-ABF8-4693-8367-D496D8F20CB8@me.com> On 8 Jul 2009, at 18:56, Wade Preston Shearer wrote: > I am submitting a form with javascript like this? > > document.formname.submit(); > > > I need the name of the form to be a variable though. Is that > possible? Meaning, that "formname" should be a variable instead of > the actual name of the form. And of course, I find the answer right after reaching out for help. eval('document.' + name + '.submit()'); -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090708/78480901/smime.bin From llihttocs at gmail.com Thu Jul 9 11:45:31 2009 From: llihttocs at gmail.com (Scott Hill) Date: Thu Jul 9 12:34:32 2009 Subject: [UPHPU] Javascript question Message-ID: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> If I have a select with several options, one of which is "selected", how do I de-select the original selected option and set a different option as selected using javascript? Is it possible? I have successfully set the values of inputs before but never this. Thanks, -- Scott Hill From bigdog at venticon.com Thu Jul 9 12:46:14 2009 From: bigdog at venticon.com (thebigdog) Date: Thu Jul 9 13:35:17 2009 Subject: [UPHPU] Javascript question In-Reply-To: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> Message-ID: <4A563AF6.3050807@venticon.com> > If I have a select with several options, one of which is "selected", how do > I de-select the original selected option and set a different option as > selected using javascript? Is it possible? I have successfully set the > values of inputs before but never this. you need to use the select and option objects with javascript and set the default. here is an example: var myselect=document.getElementById("sample") for (var i=0; i References: <4F837D6B-ADB1-44FD-B558-146C92CCEF83@bluesunhosting.com> <6360b3a10905270933m650acdd3ue96103998ed1a25e@mail.gmail.com> Message-ID: I'm a bit behind on my email and just got to this message. I just went to the link and I'm getting a 404. Is it going to be back up soon? On Wed, May 27, 2009 at 10:37 AM, Victor Villa wrote: > yipes, having the nav over the vid is bad, and no hadn't seen it. thx for > the pic! > > /me gets working on a fix > > v > From vvilla at gmail.com Thu Jul 9 20:25:07 2009 From: vvilla at gmail.com (Victor Villa) Date: Thu Jul 9 21:14:06 2009 Subject: [UPHPU] New Site Resource! In-Reply-To: References: <4F837D6B-ADB1-44FD-B558-146C92CCEF83@bluesunhosting.com> <6360b3a10905270933m650acdd3ue96103998ed1a25e@mail.gmail.com> Message-ID: hi, I'm a bit behind on my email and just got to this message. I just went to > the link and I'm getting a 404. Is it going to be back up soon? > those resources have been relocated to http://uphpu.org/category/presentations/ many thx to wps for cleaning that up and making it SEO friendly. kind feedback and critique on the videos accepted privately at vvilla@gmail.com mj/v From wadeshearer.lists at me.com Thu Jul 9 20:51:35 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Thu Jul 9 21:40:35 2009 Subject: [UPHPU] New Site Resource! In-Reply-To: References: <4F837D6B-ADB1-44FD-B558-146C92CCEF83@bluesunhosting.com> <6360b3a10905270933m650acdd3ue96103998ed1a25e@mail.gmail.com> Message-ID: On 9 Jul 2009, at 20:25, Victor Villa wrote: > those resources have been relocated to > > http://uphpu.org/category/presentations/ > > many thx to wps for cleaning that up and making it SEO friendly. > > kind feedback and critique on the videos accepted privately at > vvilla@gmail.com And many thanks to Victor for shooting and editing the video. I hope that we can update the videos with a larger size option soon as well (full screen or in a modal perhaps). The FLV player already provides a fullscreen options, so if we can export a version that is good enough quality (but still bandwidth-friendly), that be the best. -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090709/9f1c887d/smime.bin From simplepic at gmail.com Thu Jul 9 20:53:02 2009 From: simplepic at gmail.com (Alvaro Carrasco) Date: Thu Jul 9 21:42:04 2009 Subject: [UPHPU] submitting a form with javascript, dynamic form name In-Reply-To: <2E568B47-ABF8-4693-8367-D496D8F20CB8@me.com> References: <076B1073-452F-4EBB-A17D-4982F16CA774@me.com> <2E568B47-ABF8-4693-8367-D496D8F20CB8@me.com> Message-ID: <4A56AD0E.5020308@gmail.com> Hi, See below: Wade Preston Shearer wrote: > On 8 Jul 2009, at 18:56, Wade Preston Shearer wrote: > >> I am submitting a form with javascript like this? >> >> document.formname.submit(); >> >> >> I need the name of the form to be a variable though. Is that >> possible? Meaning, that "formname" should be a variable instead of >> the actual name of the form. > > And of course, I find the answer right after reaching out for help. > > eval('document.' + name + '.submit()'); > There's no need to use eval (which should be avoided when possible for various reasons). You can do this: document.forms[name].submit(); Alvaro From wadeshearer.lists at me.com Fri Jul 10 16:22:14 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Fri Jul 10 17:11:13 2009 Subject: [UPHPU] Javascript question In-Reply-To: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> Message-ID: On 9 Jul 2009, at 11:45, Scott Hill wrote: > If I have a select with several options, one of which is "selected", > how do > I de-select the original selected option and set a different option as > selected using javascript? Is it possible? I have successfully set > the > values of inputs before but never this. If you're using jQuery (which you should be), you can currently selected option in a select list like this: http://anavidesign.com/temp/select.html -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090710/fece7d66/smime.bin From wadeshearer.lists at me.com Fri Jul 10 16:27:35 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Fri Jul 10 17:16:33 2009 Subject: [UPHPU] Javascript question In-Reply-To: References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> Message-ID: On 10 Jul 2009, at 16:22, Wade Preston Shearer wrote: > If you're using jQuery (which you should be), you can currently > selected option in a select list like this: > > http://anavidesign.com/temp/select.html Sorry, that sentence was missing several words and didn't make sense. It should have read: If you're using jQuery (which you should be), you can change the currently selected option in a select list like this: -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090710/5c43bc1a/smime.bin From bigdog at venticon.com Fri Jul 10 19:25:49 2009 From: bigdog at venticon.com (thebigdog) Date: Fri Jul 10 20:14:50 2009 Subject: [UPHPU] Javascript question In-Reply-To: References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> Message-ID: <4A57EA1D.4070107@venticon.com> > If you're using jQuery (which you should be), you can change the > currently selected option in a select list like this: wow...tons more code than myselect.selectedIndex = 3; // way simple thanks, -- thebigdog From justin at justinhileman.info Sat Jul 11 07:06:58 2009 From: justin at justinhileman.info (Justin Hileman) Date: Sat Jul 11 07:56:01 2009 Subject: [UPHPU] Javascript question In-Reply-To: <4A57EA1D.4070107@venticon.com> References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> <4A57EA1D.4070107@venticon.com> Message-ID: <4A588E72.9030604@justinhileman.info> thebigdog wrote: > wow...tons more code than > > myselect.selectedIndex = 3; // way simple Don't try to break out the brevity card when you're arguing against jQuery. Try this instead: $("#sample").val(3); vs. var myselect=document.getElementById("sample"); myselect.selectedIndex = 3; -- justin http://justinhileman.com From joseph at josephscott.org Sat Jul 11 08:11:19 2009 From: joseph at josephscott.org (Joseph Scott) Date: Sat Jul 11 09:00:20 2009 Subject: [UPHPU] Javascript question In-Reply-To: <4A588E72.9030604@justinhileman.info> References: <7abc15510907091045x7d28f3c0ve479269df23ea1f5@mail.gmail.com> <4A57EA1D.4070107@venticon.com> <4A588E72.9030604@justinhileman.info> Message-ID: <7582F160-3453-4A42-A333-494CAA93BCDB@josephscott.org> On Jul 11, 2009, at 7:06 AM, Justin Hileman wrote: > Don't try to break out the brevity card when you're arguing against > jQuery. Try this instead: > > $("#sample").val(3); Oh, you forgot to include the 57,254 bytes of jQuery source above this, making it slightly bigger than the two lines you show below :-) I really like jQuery, I use it frequently and encourage others to use it often. If you are doing any amount of JS I highly recommend using it (or some JS library, jQuery being my favorite). But let's not confuse things here. This reminds of the folks who claim to write a web server 12 lines of Python (or Perl, Ruby, etc.) by conveniently not mentioning the 37,000 lines of modules that they imported and built upon. > vs. > > var myselect=document.getElementById("sample"); > myselect.selectedIndex = 3; -- Joseph Scott joseph@josephscott.org http://josephscott.org/ From jmul.php at gmail.com Sat Jul 11 16:46:58 2009 From: jmul.php at gmail.com (J M) Date: Sat Jul 11 17:35:56 2009 Subject: [UPHPU] OffTopic: B-day cake... Message-ID: ...fit for a programmer. (for those that didn't catch the link off of digg.com) http://www.thedoghousediaries.com/comics/uncategorized/2009-06-05-e38e6cf824a0fac7718d7d99a804fbb7.png From miles.togoe at gmail.com Sat Jul 11 18:20:04 2009 From: miles.togoe at gmail.com (MilesTogoe) Date: Sat Jul 11 19:10:09 2009 Subject: [UPHPU] tracking domain calls Message-ID: <4A592C34.1010206@gmail.com> I switched one of our domains to GoDaddy, pointing it to a linode vps. But the domain just hangs in space. The godaddy account seems to be pointed to the correct linode servers (NS1.LINODE.COM, ...). The apache virtual domain host on our linode account seems to be right: sudo vim /etc/apache2/sites-available/ourdomain ServerName ourdomain.com ServerAlias www.ourdomain.com DocumentRoot /home/ourdomain/public sudo a2ensite ourdomain sudo /etc/init.d/apache2 reload but all I get is nothing - no errors, just nothing. So is there any way to trace the browser call - to see where it is trying to go ? So I can tell if the error is on the godaddy side or the linode side ? thks for any help. From haas at xmission.com Sat Jul 11 20:30:49 2009 From: haas at xmission.com (Walt Haas) Date: Sat Jul 11 21:20:09 2009 Subject: [UPHPU] tracking domain calls In-Reply-To: <4A592C34.1010206@gmail.com> References: <4A592C34.1010206@gmail.com> Message-ID: <4A594AD9.7080908@xmission.com> MilesTogoe wrote: > I switched one of our domains to GoDaddy, pointing it to a linode vps. > But the domain just hangs in space. The godaddy account seems to be > pointed to the correct linode servers (NS1.LINODE.COM, ...). The apache > virtual domain host on our linode account seems to be right: > > sudo vim /etc/apache2/sites-available/ourdomain > > ServerName ourdomain.com > ServerAlias www.ourdomain.com > DocumentRoot /home/ourdomain/public > > sudo a2ensite ourdomain > sudo /etc/init.d/apache2 reload > > but all I get is nothing - no errors, just nothing. So is there any way > to trace the browser call - to see where it is trying to go ? So I can > tell if the error is on the godaddy side or the linode side ? > thks for any help. I'm already lost. What does GoDaddy resolve for you? An A record or a domain delegation? -- Walt From wadeshearer.lists at me.com Sat Jul 11 20:34:44 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Sat Jul 11 21:23:41 2009 Subject: [UPHPU] tracking domain calls In-Reply-To: <4A592C34.1010206@gmail.com> References: <4A592C34.1010206@gmail.com> Message-ID: <672A2E42-2CFD-4B32-B3C5-2BA960755097@me.com> On 11 Jul 2009, at 18:20, MilesTogoe wrote: > I switched one of our domains to GoDaddy, pointing it to a linode > vps. But the domain just hangs in space. The godaddy account seems > to be pointed to the correct linode servers (NS1.LINODE.COM, ...). > The apache virtual domain host on our linode account seems to be > right: > > but all I get is nothing - no errors, just nothing. So is there any > way to trace the browser call - to see where it is trying to go ? > So I can tell if the error is on the godaddy side or the linode side ? > thks for any help. What do you get when you ping the domain? -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090711/44124116/smime.bin From lists at sollis.net Sun Jul 12 08:49:48 2009 From: lists at sollis.net (Chad Sollis) Date: Sun Jul 12 09:37:37 2009 Subject: [UPHPU] tracking domain calls In-Reply-To: <672A2E42-2CFD-4B32-B3C5-2BA960755097@me.com> References: <4A592C34.1010206@gmail.com> <672A2E42-2CFD-4B32-B3C5-2BA960755097@me.com> Message-ID: <86403C58-5FB6-426E-96CC-36E234DC0AEF@sollis.net> Does linode act exclusively as a vps, or does it have name servers as well? If linode does have name servers, then your domain needs to be configured to route *.domain.com to your vps IP Address from their name servers. If linode does not have name servers you can do one of two things, 1) use godaddy name servers, and then use total dns control in godaddy to point to your vps IP address. 2) setup an account with everydns.net (or similar), change your godaddy name server settings to NS1.EVERYDNS.NET(or similar) and then from everydns.net route your domain to the vps IP address. Cheers ~SOL On Jul 11, 2009, at 8:34 PM, Wade Preston Shearer wrote: > On 11 Jul 2009, at 18:20, MilesTogoe wrote: > >> I switched one of our domains to GoDaddy, pointing it to a linode >> vps. But the domain just hangs in space. The godaddy account >> seems to be pointed to the correct linode servers >> (NS1.LINODE.COM, ...). The apache virtual domain host on our >> linode account seems to be right: >> >> but all I get is nothing - no errors, just nothing. So is there >> any way to trace the browser call - to see where it is trying to >> go ? So I can tell if the error is on the godaddy side or the >> linode side ? >> thks for any help. > > What do you get when you ping the domain? > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net From miles.togoe at gmail.com Sun Jul 12 12:21:26 2009 From: miles.togoe at gmail.com (MilesTogoe) Date: Sun Jul 12 13:11:16 2009 Subject: [UPHPU] tracking domain calls In-Reply-To: <86403C58-5FB6-426E-96CC-36E234DC0AEF@sollis.net> References: <4A592C34.1010206@gmail.com> <672A2E42-2CFD-4B32-B3C5-2BA960755097@me.com> <86403C58-5FB6-426E-96CC-36E234DC0AEF@sollis.net> Message-ID: <4A5A29A6.5030501@gmail.com> On 07/12/2009 08:49 AM, Chad Sollis wrote: > Does linode act exclusively as a vps, or does it have name servers as > well? > > If linode does have name servers, then your domain needs to be > configured to route *.domain.com to your vps IP Address from their > name servers. okay - problem solved. Thanks for tip Chris - yes, there is a domain manager step in linode that I had skipped to add the domain to our account. and thks Wade for the ping suggestion as a good check (should have been obvious but I forgot it). Now on to routing mail to google apps (from suggestions in this forum). > > If linode does not have name servers you can do one of two things, > > 1) use godaddy name servers, and then use total dns control in godaddy > to point to your vps IP address. > > 2) setup an account with everydns.net (or similar), change your > godaddy name server settings to NS1.EVERYDNS.NET(or similar) and then > from everydns.net route your domain to the vps IP address. > > Cheers > ~SOL > > > On Jul 11, 2009, at 8:34 PM, Wade Preston Shearer wrote: > >> On 11 Jul 2009, at 18:20, MilesTogoe wrote: >> >>> I switched one of our domains to GoDaddy, pointing it to a linode >>> vps. But the domain just hangs in space. The godaddy account seems >>> to be pointed to the correct linode servers (NS1.LINODE.COM, ...). >>> The apache virtual domain host on our linode account seems to be right: >>> >>> but all I get is nothing - no errors, just nothing. So is there any >>> way to trace the browser call - to see where it is trying to go ? >>> So I can tell if the error is on the godaddy side or the linode side ? >>> thks for any help. >> >> What do you get when you ping the domain? >> >> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From vvilla at gmail.com Mon Jul 13 12:42:48 2009 From: vvilla at gmail.com (Victor Villa) Date: Mon Jul 13 13:31:41 2009 Subject: [UPHPU] Meeting This Thursday 7p - Josh Fenio - A Web Developer's Guide to System Security Message-ID: Greets All! I'm pleased to announce this Thursday's meeting A Web Developer's Guide to System Security. The presentation will track the day to day aspects of a security-minded systems administrator. Topics that will be covered include intrusion prevention, intrusion detection, and recovery, along with relevant tools to accomplish each mission. Josh Fenio is a software engineer in Northern Nevada, known as "stderr" or "dataw0lf" on the IRCs. mj/v From lists at sollis.net Thu Jul 16 09:38:09 2009 From: lists at sollis.net (Chad Sollis) Date: Thu Jul 16 09:46:46 2009 Subject: [UPHPU] Cross domain javascript request Message-ID: <8c9aad0907160838m53383c82w113bf78d4adb28f5@mail.gmail.com> Greetings, I know this topic has been covered in the past, so forgive the repeat. Looking for some quick feedback/ideas. I need to make a call from *a*.domain.com to *b*.domain.com "behind the scenes" but I do not need to read a response, just make the POST/GET request. >From the quick search I have done, I have found a couple of options. I am wondering if there is a better option, or of the options below, what is the consesous on which is best. Thanks in advance for your help. 1. use an iframe with height and width of 0 or 1 to pass the data to b.domain.com 2. use jquery $.getJSON() it will make the request cross domain, I simply won't return anything. 3. use a proxy on a.domain.com I actually prefer #1, because the code I use is included in many place, some with jquery, some without jquery, and with option #3, I dont know if I have access to create a proxy on that server. Is iframe a bad route? thanks all. ~SOL From vvilla at gmail.com Thu Jul 16 10:12:38 2009 From: vvilla at gmail.com (Victor Villa) Date: Thu Jul 16 10:21:17 2009 Subject: [UPHPU] Meeting Today at 7p - Josh Fenio - A Web Developer's Guide to System Security Message-ID: Greets All! I'm pleased to announce this Thursday's meeting A Web Developer's Guide to System Security. The presentation will track the day to day aspects of a security-minded systems administrator. Topics that will be covered include intrusion prevention, intrusion detection, and recovery, along with relevant tools to accomplish each mission. Josh Fenio is a software engineer in Northern Nevada, known as "stderr" or "dataw0lf" on the IRCs. Meeting Info A Web Developer's Guide to System Security Josh Fenio July 16, 2009 at 7pm Meeting at Bill Good Marketing 12393 Gateway Park Place Suite 600 Draper, Ut 84020 mj/v From wadeshearer.lists at me.com Thu Jul 16 10:47:13 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Thu Jul 16 10:57:09 2009 Subject: [UPHPU] Cross domain javascript request In-Reply-To: <8c9aad0907160838m53383c82w113bf78d4adb28f5@mail.gmail.com> References: <8c9aad0907160838m53383c82w113bf78d4adb28f5@mail.gmail.com> Message-ID: <164842065405195607515344071332545951391-Webmail@me.com> On Thursday, July 16, 2009, at 08:38AM, "Chad Sollis" wrote: >I need to make a call from *a*.domain.com to *b*.domain.com "behind the >scenes" but I do not need to read a response, just make the POST/GET >request. > >>From the quick search I have done, I have found a couple of options. I am >wondering if there is a better option, or of the options below, what is the >consesous on which is best. > >Thanks in advance for your help. > > > 1. use an iframe with height and width of 0 or 1 to pass the data to > b.domain.com > 2. use jquery $.getJSON() it will make the request cross domain, I simply > won't return anything. > 3. use a proxy on a.domain.com Javascript would be my preference, but no, I don't think that an iframe is an unacceptable solution. It seems like a hack, but if it gets it done, it gets it done. From kirk at kirkouimet.com Tue Jul 21 08:53:42 2009 From: kirk at kirkouimet.com (Kirk Ouimet) Date: Tue Jul 21 09:02:39 2009 Subject: [UPHPU] A Good GUID/UUID PHP Solution Message-ID: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> Hi List, Can anyone recommend a good PHP function/class to generate GUIDs? I'm shifting from using auto-increment to using GUIDs - any advice for this type of change? Does a SQL lookup slow down using a GUID as a PK versus using an integer as a PK? Thanks! Kirk Ouimet kirk@kirkouimet.com Cell: (801) 310-1421 From kirk at kirkouimet.com Tue Jul 21 09:04:23 2009 From: kirk at kirkouimet.com (Kirk Ouimet) Date: Tue Jul 21 09:13:16 2009 Subject: [UPHPU] Should I Be Using an ORM? Message-ID: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> Hi List, Had a couple questions to post this morning. This one is very open-ended. I've been looking at PHP ORM's like: http://www.doctrine-project.org/ And am curious to know if it's worth my time to learn and use a system like this. Thoughts? Kirk Ouimet kirk@kirkouimet.com Cell: (801) 310-1421 From jgiboney at gmail.com Tue Jul 21 09:10:28 2009 From: jgiboney at gmail.com (Justin Giboney) Date: Tue Jul 21 09:19:03 2009 Subject: [UPHPU] A Good GUID/UUID PHP Solution In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> References: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> Message-ID: Most databases have a built in GUID/UUID generator http://dev.mysql.com/doc/refman/5.0/en/miscellaneous-functions.html#function_uuid Just yesterday I was looking into doing the same thing. I decided against it because of the speed impact that you will have on your database. I was talking to a friend at a large high-tech company that has terrabytes and they say that they use auto-increment for performance issues. Good luck with whatever you decide. Justin Giboney On Jul 21, 2009, at 7:53 AM, Kirk Ouimet wrote: > Hi List, > > > > Can anyone recommend a good PHP function/class to generate GUIDs? I'm > shifting from using auto-increment to using GUIDs - any advice for > this type > of change? Does a SQL lookup slow down using a GUID as a PK versus > using an > integer as a PK? > > > > Thanks! > > > > Kirk Ouimet > > kirk@kirkouimet.com > > Cell: (801) 310-1421 > > > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net From kirksemail at gmail.com Tue Jul 21 09:44:27 2009 From: kirksemail at gmail.com (Kirk Cerny) Date: Tue Jul 21 09:52:57 2009 Subject: [UPHPU] A Good GUID/UUID PHP Solution In-Reply-To: References: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> Message-ID: <7037ffd10907210844m61cc7dc8p3eaf94515971bf49@mail.gmail.com> I run a small site where I use GUID's along side auto_increment id's. The GUID is only used by the browser or client and the php uses the real id. http://72.8.94.16/user/addMember/index.html Look at the code for the state select box, the values are really hashed numbers. I am able to turn the hash back into a number because I just use the mysql AES_ENCRYPT / AES_DECRYPT methods. I have the hash list cached so I do not run the encryption queries on every page load. I still look up the state by its auto_increment id in the states table. I would like feedback on whether this is a good idea or not. Kirk Cerny On Tue, Jul 21, 2009 at 9:10 AM, Justin Giboney wrote: > Most databases have a built in GUID/UUID generator > > http://dev.mysql.com/doc/refman/5.0/en/miscellaneous-functions.html#function_uuid > > Just yesterday I was looking into doing the same thing. I decided against it > because of the speed impact that you will have on your database. > > I was talking to a friend at a large high-tech company that has terrabytes > and they say that they use auto-increment for performance issues. > > Good luck with whatever you decide. > > Justin Giboney > > > On Jul 21, 2009, at 7:53 AM, Kirk Ouimet wrote: > >> Hi List, >> >> >> >> Can anyone recommend a good PHP function/class to generate GUIDs? I'm >> shifting from using auto-increment to using GUIDs - any advice for this >> type >> of change? Does a SQL lookup slow down using a GUID as a PK versus using >> an >> integer as a PK? >> >> >> >> Thanks! >> >> >> >> Kirk Ouimet >> >> kirk@kirkouimet.com >> >> Cell: (801) 310-1421 >> >> >> >> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From simplepic at gmail.com Tue Jul 21 11:24:54 2009 From: simplepic at gmail.com (Alvaro Carrasco) Date: Tue Jul 21 11:33:30 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> Message-ID: <4A65F9E6.9040000@gmail.com> Kirk Ouimet wrote: > Hi List, > > > > Had a couple questions to post this morning. This one is very open-ended. > I've been looking at PHP ORM's like: > > > > http://www.doctrine-project.org/ > > > > And am curious to know if it's worth my time to learn and use a system like > this. Thoughts? > > > > Kirk Ouimet > > kirk@kirkouimet.com > > Cell: (801) 310-1421 > > I think the main reason for an ORM is to allow you to more easily create a Domain Model [PoEA]. If all you're doing is managing rows in the database or creating reports, then you probably don't need one. If you're looking to create a system that captures the complexities of a domain in a way that is easy to understand, maintain, and extend then you might want to model the domain that you're working on using objects and take advantage of the the many patterns available when using object-oriented programming. A nice ORM will help you head in that direction. I personally use an ORM for almost all of my projects. I personally think it's generally more enjoyable to work with a nice ORM than without. I have used Propel in the past and have looked at Doctrine. I think the coolest one by far is Outlet http://www.outlet-orm.org (a little biased, I wrote it :P ). It's much smaller than either Propel or Doctrine but it provides a LOT of nice features while staying very transparent. Keep in mind that you'll likely still be making sql calls here and there and that's fine. An ORM is ill-suited for doing reports, batch updates, aggregations, etc. Alvaro From gthornock at yahoo.com Tue Jul 21 12:20:58 2009 From: gthornock at yahoo.com (Gary Thornock) Date: Tue Jul 21 12:29:26 2009 Subject: [UPHPU] A Good GUID/UUID PHP Solution In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> Message-ID: <67839.61932.qm@web30006.mail.mud.yahoo.com> --- On Tue, 7/21/09, Kirk Ouimet wrote: > Can anyone recommend a good PHP function/class to generate > GUIDs? I'm shifting from using auto-increment to using GUIDs - > any advice for this type of change? Does a SQL lookup slow down > using a GUID as a PK versus using an integer as a PK? Database engines that use clustered indexes (data rows whose primary keys are close in value are physically stored close together on disk) will see a slowdown if you use any randomly distributed value (including a GUID) as the primary key. Inserts are particularly bad, but reads can be problematic as well because there's no locality to help in caching "nearby" values. GUIDs can be useful as your application's key (which you would set as an additional unique index on the table), but for the database engine, you're usually better off using a consistently increasing value like an auto-increment as the primary key. Note, of course, that "usually" != "always", you may find some particular edge case where the rule of thumb doesn't apply. From bigdog at venticon.com Tue Jul 21 15:39:12 2009 From: bigdog at venticon.com (thebigdog) Date: Tue Jul 21 15:47:43 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> Message-ID: <4A663580.9060706@venticon.com> > Had a couple questions to post this morning. This one is very open-ended. > I've been looking at PHP ORM's like: > > http://www.doctrine-project.org/ > > And am curious to know if it's worth my time to learn and use a system like > this. Thoughts? ORM's are pretty good to have if you dont want to spend the time writing all the database backend code. It can be a life-saver on many projects. If you are not familiar with the concepts that ORM offers especially with PHP or the language that you are writing in, then I would suggest spending the time to learn the key concepts and theory behind it first, so that you have a solid foundation. Then from there you can make an educated decision on the ORM project that you think would be beneficial for your project. I have ran into issues on numerous projects where the ORM tools where decided without ample review of the business logic or design - thus hampering the project. Also if you are not up to par on PHP Object design methodology it might be good to do that. Also I would look at which database you are using so you know the limitations of ORM with your specific db. (Sometimes there are better ways to do things like mass deletes without going through ORM that you might need to be aware of) With that said, it is very ideal to map your db to objects. hth, -- thebigdog From miles.togoe at gmail.com Tue Jul 21 16:00:12 2009 From: miles.togoe at gmail.com (MilesTogoe) Date: Tue Jul 21 16:09:51 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <4A663580.9060706@venticon.com> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> <4A663580.9060706@venticon.com> Message-ID: <4A663A6C.1060706@gmail.com> On 07/21/2009 03:39 PM, thebigdog wrote: >> Had a couple questions to post this morning. This one is very open-ended. >> I've been looking at PHP ORM's like: >> >> http://www.doctrine-project.org/ >> >> And am curious to know if it's worth my time to learn and use a system like >> this. Thoughts? >> > > ORM's are pretty good to have if you dont want to spend the time writing all the > database backend code. It can be a life-saver on many projects. If you are not > familiar with the concepts that ORM offers especially with PHP or the language > that you are writing in, then I would suggest spending the time to learn the key > concepts and theory behind it first, so that you have a solid foundation. Then > from there you can make an educated decision on the ORM project that you think > would be beneficial for your project. I have ran into issues on numerous > projects where the ORM tools where decided without ample review of the business > logic or design - thus hampering the project. Also if you are not up to par on > PHP Object design methodology it might be good to do that. Also I would look at > which database you are using so you know the limitations of ORM with your > specific db. (Sometimes there are better ways to do things like mass deletes > without going through ORM that you might need to be aware of) With that said, it > is very ideal to map your db to objects. > unless of course if you plan to use one of the "key:value" databases (often found in cloud db's) These would include amazon, google big table, couchdb, tokyo cabinet, mongodb, ......... These allow you to store your data in object form and then make queries directly without SQL. From justin at justinhileman.info Tue Jul 21 20:10:25 2009 From: justin at justinhileman.info (Justin Hileman) Date: Tue Jul 21 20:18:56 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> Message-ID: <4A667511.8000106@justinhileman.info> Kirk Ouimet wrote: > Had a couple questions to post this morning. This one is very open-ended. > I've been looking at PHP ORM's like: > > > > http://www.doctrine-project.org/ They're absolutely worth your time to learn. They're super rad, and usually a bit fickle. And, as others have pointed out, they're not the best answer for everything. But they're definitely the best answer for a lot of things. My team is using Doctrine on a PHP platform with a pretty complicated model, and the ORM makes traversing that model *far* easier. Unfortunately, Doctrine isn't magic. One of my guys threw together a bulk import script that could peg a dual-quad Xeon until PHP timed out and still not finish... We converted the inner loop into a direct SQL insert and the import script could finish in a couple of minutes. So be sure to watch for places you can optimize. That said, we still use the ORM for almost everything. We have over a hundred model classes in in the project (~150 DB tables) and we have about 18 "normal" SQL queries in the app. I was curious so I just grepped the source code. Looks like we fall back to SQL in only a couple of cases: * Bulk import, as mentioned above * Bulk export/reporting -- although a lot of reporting is done through Doctrine, because it builds all the heinous joins for us :) * Pagination -- The version of Doctrine we're using has a bug that misreports pagination data... It pages the queries just fine, but lies about how many total rows exist, so we ignore what it tells us and ask directly via SQL. * And one other place where I'm pretty sure it should have been done in Doctrine. I'll chat with the developer that wrote that code, see if he has a good reason for it :) -- justin http://justinhileman.com From justin at justinhileman.info Tue Jul 21 20:19:11 2009 From: justin at justinhileman.info (Justin Hileman) Date: Tue Jul 21 20:27:43 2009 Subject: [UPHPU] A Good GUID/UUID PHP Solution In-Reply-To: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> References: <75DEA8DC63249446B84C2133FDC75A3827290A5A@VMBX115.ihostexchange.net> Message-ID: <4A66771F.3060101@justinhileman.info> Kirk Ouimet wrote: > Can anyone recommend a good PHP function/class to generate GUIDs? com_create_guid() works if you've got it... -- justin http://justinhileman.com From cboyack at gmail.com Tue Jul 21 23:02:10 2009 From: cboyack at gmail.com (Connor Boyack) Date: Tue Jul 21 23:10:58 2009 Subject: [UPHPU] Cross domain javascript request In-Reply-To: <8c9aad0907160838m53383c82w113bf78d4adb28f5@mail.gmail.com> References: <8c9aad0907160838m53383c82w113bf78d4adb28f5@mail.gmail.com> Message-ID: <91a4f7510907212202l720454cfs136c603525e7e24b@mail.gmail.com> > 3. use a proxy on a.domain.com > This is the solution I use and prefer (a PHP proxy using CURL) since it can be abstracted to fetch other URLs as well. I just create a base proxy that can accept certain parameters and then I can reuse that as necessary on other domains, projects, etc. Connor -- Connor Boyack connorboyack.com "The true source of our suffering has been our timidity. We have been afraid to think. . . . Let us dare to read, think, speak, and write." ?John Adams Join my email list: http://tinyurl.com/ywx5lk From lists at sollis.net Wed Jul 22 10:53:05 2009 From: lists at sollis.net (Chad Sollis) Date: Wed Jul 22 11:01:32 2009 Subject: [UPHPU] looking for opinions/ideas Message-ID: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> Greetings, I am scoping a new enterprise grade web application. I am simply in research mode right now, but interested in any feedback from the group as it relates to ideas on technology, scale, cost efficiency, development, performance, etc, in addition to creating an environment that talented engineers could enter and find their way around very easily. I am planning on using PHP for a variety of reasons using AOSD/SOA style of architecture. I am open to other suggestions if the consensus indicates that is not the right route. I know this is kind of broad question, if you have experience in even one of the areas, any feedback is appreciated. Ideas under consideration: - Frameworks / ORM (performance, scale, and customization): - Zend - Cake - Codeigniter - Doctrine - Project - Propel - Database (triggers, procedures, speed and scale) - MySQL - Postgres - Netezza - Amazon SimpleDB - Architecture (scale) - EC2 - Mosso - Terremark - Managed Servers Many Many Thanks for your ideas and suggestions. ~SOL From utahphp at forsalesticker.com Wed Jul 22 11:32:07 2009 From: utahphp at forsalesticker.com (CarSign) Date: Wed Jul 22 11:40:34 2009 Subject: [UPHPU] looking for opinions/ideas Message-ID: <221452.53728.qm@web52011.mail.re2.yahoo.com> I will comment on the database part of your app. I am in favor of using stored procedures for the following reasons (if they apply to your web app) They provide a greater degree of security between your web app and the DB. (you can restrict the web applications communication to the DB to be only through stored procedures) Can run one procedure and get multiple results. allows more logic to be performed on the DB which can reduce load on PHP. Can provide a level of abstraction for your programmers. (So one or two of your programmers can be more involved with the specifics of the DB. Such as requiring that the design is correct, calls are optimized and such. Where your other programmers can focus more on the web app.) On a side note: I know that some of the frameworks out there have DB objects codded into them (specificaly I tried SPs in CakePHP) and they may play that nicely with using Stored Procedures. Good luck. Andrew Cluff --- On Wed, 7/22/09, Chad Sollis wrote: > From: Chad Sollis > Subject: [UPHPU] looking for opinions/ideas > To: uphpu@uphpu.org > Date: Wednesday, July 22, 2009, 10:53 AM > Greetings, > > I am scoping a new enterprise grade web application.? > I am simply in > research mode right now, but interested in any feedback > from the group as it > relates to ideas on technology, scale, cost efficiency, > development, > performance, etc, in addition to creating an environment > that talented > engineers could enter and find their way around very > easily. > > I am planning on using PHP for a variety of reasons using > AOSD/SOA style of > architecture.???I am open to other > suggestions if the consensus indicates > that is not the right route. > > I know this is kind of broad question, if you have > experience in even one of > the areas, any feedback is appreciated. > > Ideas under consideration: > > ???- Frameworks / ORM (performance, scale, > and customization): > ? ? ? - Zend > ? ? ? - Cake > ? ? ? - Codeigniter > ???- Doctrine - Project > ? ? ? - Propel > ???- Database (triggers, procedures, speed > and scale) > ???- MySQL > ? ? ? - Postgres > ? ? ? - Netezza > ? ? ? - Amazon SimpleDB > ???- Architecture (scale) > ???- EC2 > ? ? ? - Mosso > ? ? ? - Terremark > ? ? ? - Managed Servers > > Many Many Thanks for your ideas and suggestions. > > ~SOL > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From bms at mscis.org Wed Jul 22 11:36:45 2009 From: bms at mscis.org (Brandon Stout) Date: Wed Jul 22 11:45:47 2009 Subject: [UPHPU] Funeral announcement Message-ID: <4A674E2D.2020605@mscis.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 For those of you who know me who have not already heard, my wife Julie's two-year battle with a cancerous brain tumor ended last Sunday. I wrote about her passing here: http://flfn.org/blog/brasto/julie-stout-free And I posted the obituary here: http://flfn.org/blog/brasto/julies-obituary I've been busy caring for her, and thus less involved in this channel over the past few months. Soon, I'll likely be a more regular participant again, and I'm looking forward to it. Thank you to all of you who have shown your support to us. Brandon -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iEYEARECAAYFAkpnTi0ACgkQx0pgn74qrcI+IQCfS3geq+c/OsTqQNm1hK7gRSuV wsAAoL5U3wWG7DVKYgydZB8F4UUnyLoG =mGIX -----END PGP SIGNATURE----- From haas at xmission.com Wed Jul 22 11:52:35 2009 From: haas at xmission.com (Walt Haas) Date: Wed Jul 22 12:01:17 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> References: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> Message-ID: <4A6751E3.2080400@xmission.com> We tend to start with Drupal and develop any necessary custom modules. -- Walt Chad Sollis wrote: > Greetings, > > I am scoping a new enterprise grade web application. I am simply in > research mode right now, but interested in any feedback from the group as it > relates to ideas on technology, scale, cost efficiency, development, > performance, etc, in addition to creating an environment that talented > engineers could enter and find their way around very easily. > > I am planning on using PHP for a variety of reasons using AOSD/SOA style of > architecture. I am open to other suggestions if the consensus indicates > that is not the right route. > > I know this is kind of broad question, if you have experience in even one of > the areas, any feedback is appreciated. > > Ideas under consideration: > > - Frameworks / ORM (performance, scale, and customization): > - Zend > - Cake > - Codeigniter > - Doctrine - Project > - Propel > - Database (triggers, procedures, speed and scale) > - MySQL > - Postgres > - Netezza > - Amazon SimpleDB > - Architecture (scale) > - EC2 > - Mosso > - Terremark > - Managed Servers > > Many Many Thanks for your ideas and suggestions. > > ~SOL > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From lists at sollis.net Wed Jul 22 11:58:30 2009 From: lists at sollis.net (Chad Sollis) Date: Wed Jul 22 12:06:57 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: <4A6751E3.2080400@xmission.com> References: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> <4A6751E3.2080400@xmission.com> Message-ID: <8c9aad0907221058l6cb1d0au15e25fac0da3fe6d@mail.gmail.com> Are there any licensing issues if the application is commercial? On Wed, Jul 22, 2009 at 11:52 AM, Walt Haas wrote: > We tend to start with Drupal and develop any necessary custom modules. > > -- Walt > > Chad Sollis wrote: > >> Greetings, >> >> I am scoping a new enterprise grade web application. I am simply in >> research mode right now, but interested in any feedback from the group as >> it >> relates to ideas on technology, scale, cost efficiency, development, >> performance, etc, in addition to creating an environment that talented >> engineers could enter and find their way around very easily. >> >> I am planning on using PHP for a variety of reasons using AOSD/SOA style >> of >> architecture. I am open to other suggestions if the consensus indicates >> that is not the right route. >> >> I know this is kind of broad question, if you have experience in even one >> of >> the areas, any feedback is appreciated. >> >> Ideas under consideration: >> >> - Frameworks / ORM (performance, scale, and customization): >> - Zend >> - Cake >> - Codeigniter >> - Doctrine - Project >> - Propel >> - Database (triggers, procedures, speed and scale) >> - MySQL >> - Postgres >> - Netezza >> - Amazon SimpleDB >> - Architecture (scale) >> - EC2 >> - Mosso >> - Terremark >> - Managed Servers >> >> Many Many Thanks for your ideas and suggestions. >> >> ~SOL >> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net >> >> > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From haas at xmission.com Wed Jul 22 12:08:29 2009 From: haas at xmission.com (Walt Haas) Date: Wed Jul 22 12:16:56 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: <8c9aad0907221058l6cb1d0au15e25fac0da3fe6d@mail.gmail.com> References: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> <4A6751E3.2080400@xmission.com> <8c9aad0907221058l6cb1d0au15e25fac0da3fe6d@mail.gmail.com> Message-ID: <4A67559D.8030001@xmission.com> Drupal is GPL code. It can be used commercially but IANAL and don't know how it would work legally if you wrote a proprietary module that you didn't want to distribute. -- Walt Chad Sollis wrote: > Are there any licensing issues if the application is commercial? > > On Wed, Jul 22, 2009 at 11:52 AM, Walt Haas wrote: > >> We tend to start with Drupal and develop any necessary custom modules. >> >> -- Walt >> >> Chad Sollis wrote: >> >>> Greetings, >>> >>> I am scoping a new enterprise grade web application. I am simply in >>> research mode right now, but interested in any feedback from the group as >>> it >>> relates to ideas on technology, scale, cost efficiency, development, >>> performance, etc, in addition to creating an environment that talented >>> engineers could enter and find their way around very easily. >>> >>> I am planning on using PHP for a variety of reasons using AOSD/SOA style >>> of >>> architecture. I am open to other suggestions if the consensus indicates >>> that is not the right route. >>> >>> I know this is kind of broad question, if you have experience in even one >>> of >>> the areas, any feedback is appreciated. >>> >>> Ideas under consideration: >>> >>> - Frameworks / ORM (performance, scale, and customization): >>> - Zend >>> - Cake >>> - Codeigniter >>> - Doctrine - Project >>> - Propel >>> - Database (triggers, procedures, speed and scale) >>> - MySQL >>> - Postgres >>> - Netezza >>> - Amazon SimpleDB >>> - Architecture (scale) >>> - EC2 >>> - Mosso >>> - Terremark >>> - Managed Servers >>> >>> Many Many Thanks for your ideas and suggestions. >>> >>> ~SOL >>> >>> _______________________________________________ >>> >>> UPHPU mailing list >>> UPHPU@uphpu.org >>> http://uphpu.org/mailman/listinfo/uphpu >>> IRC: #uphpu on irc.freenode.net >>> >>> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net >> > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From joseph at josephscott.org Wed Jul 22 12:36:55 2009 From: joseph at josephscott.org (Joseph Scott) Date: Wed Jul 22 12:45:29 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: <221452.53728.qm@web52011.mail.re2.yahoo.com> References: <221452.53728.qm@web52011.mail.re2.yahoo.com> Message-ID: On Jul 22, 2009, at 11:32 AM, CarSign wrote: > allows more logic to be performed on the DB which can reduce load on > PHP. I'd suggest that this can be both a good thing and bad thing depending on the needs of the project. If the project requirements are such that it has to scale to large number of simultaneous users and page views then this would be a bad thing. It's much easier to spread the load at the web server/PHP level than at the database level. On the other side, if this app is targeted for most controlled environments like internal business deployments, then pushing more logic into the DB can be very helpful. You're often less concerned about scale since you generally have an upper bound on the number of concurrent users (number of employees). And having more logic in the DB can make it easy for other applications to speak to the DB without breaking things (that might be a non-issue if provide a really good API that other apps can use). My current job demands using the first approach, my previous job we did apps that used the second. It varies depending on the needs, requirements, expectations and limits of your project. -- Joseph Scott joseph@josephscott.org http://josephscott.org/ From jgiboney at gmail.com Thu Jul 23 00:35:52 2009 From: jgiboney at gmail.com (Justin Giboney) Date: Thu Jul 23 00:44:21 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <4A667511.8000106@justinhileman.info> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> <4A667511.8000106@justinhileman.info> Message-ID: <990870EB-10AD-416F-9D8B-E5D13D885BA4@gmail.com> To follow up on this question, can these ORMs be installed if you don't have access to the full server. For example I have a client that setup his host as Host Monster and I don't know if I'll be able to install an ORM on it. Doctrine seems to need to be installed as an application. Are there any that don't require full access to the server? Thanks, Justin Giboney On Jul 21, 2009, at 7:10 PM, Justin Hileman wrote: > Kirk Ouimet wrote: >> Had a couple questions to post this morning. This one is very open- >> ended. >> I've been looking at PHP ORM's like: >> >> >> >> http://www.doctrine-project.org/ > > > > They're absolutely worth your time to learn. They're super rad, and > usually a bit fickle. And, as others have pointed out, they're not > the best answer for everything. But they're definitely the best > answer for a lot of things. > > My team is using Doctrine on a PHP platform with a pretty > complicated model, and the ORM makes traversing that model *far* > easier. Unfortunately, Doctrine isn't magic. One of my guys threw > together a bulk import script that could peg a dual-quad Xeon until > PHP timed out and still not finish... We converted the inner loop > into a direct SQL insert and the import script could finish in a > couple of minutes. So be sure to watch for places you can optimize. > > That said, we still use the ORM for almost everything. We have over > a hundred model classes in in the project (~150 DB tables) and we > have about 18 "normal" SQL queries in the app. > > I was curious so I just grepped the source code. Looks like we fall > back to SQL in only a couple of cases: > > * Bulk import, as mentioned above > > * Bulk export/reporting -- although a lot of reporting is done > through Doctrine, because it builds all the heinous joins for us :) > > * Pagination -- The version of Doctrine we're using has a bug that > misreports pagination data... It pages the queries just fine, but > lies about how many total rows exist, so we ignore what it tells us > and ask directly via SQL. > > * And one other place where I'm pretty sure it should have been done > in Doctrine. I'll chat with the developer that wrote that code, see > if he has a good reason for it :) > > > -- > justin > http://justinhileman.com > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net From haas at xmission.com Thu Jul 23 05:43:29 2009 From: haas at xmission.com (Walt Haas) Date: Thu Jul 23 05:51:56 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <990870EB-10AD-416F-9D8B-E5D13D885BA4@gmail.com> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> <4A667511.8000106@justinhileman.info> <990870EB-10AD-416F-9D8B-E5D13D885BA4@gmail.com> Message-ID: <4A684CE1.5010703@xmission.com> Well, from http://www.doctrine-project.org/documentation/manual/1_1/en/introduction#minimum-requirements you would need PHP 5.2.3 or later and the PDO extension, and from http://www.doctrine-project.org/documentation/cookbook/1_1/en/my-first-project you need command line access to the server. According to this http://helpdesk.hostmonster.com/index.php/kb/article/000179 Host Monster does give you SSH access on request. But the only way I see to find out what PHP version is available is to run phpinfo() on the server, and you need to log in to the account to do that. -- Walt Justin Giboney wrote: > To follow up on this question, can these ORMs be installed if you don't > have access to the full server. For example I have a client that setup > his host as Host Monster and I don't know if I'll be able to install an > ORM on it. Doctrine seems to need to be installed as an application. > > Are there any that don't require full access to the server? > > Thanks, > > Justin Giboney > > > On Jul 21, 2009, at 7:10 PM, Justin Hileman wrote: > >> Kirk Ouimet wrote: >>> Had a couple questions to post this morning. This one is very >>> open-ended. >>> I've been looking at PHP ORM's like: >>> >>> >>> >>> http://www.doctrine-project.org/ >> >> >> >> They're absolutely worth your time to learn. They're super rad, and >> usually a bit fickle. And, as others have pointed out, they're not the >> best answer for everything. But they're definitely the best answer for >> a lot of things. >> >> My team is using Doctrine on a PHP platform with a pretty complicated >> model, and the ORM makes traversing that model *far* easier. >> Unfortunately, Doctrine isn't magic. One of my guys threw together a >> bulk import script that could peg a dual-quad Xeon until PHP timed out >> and still not finish... We converted the inner loop into a direct SQL >> insert and the import script could finish in a couple of minutes. So >> be sure to watch for places you can optimize. >> >> That said, we still use the ORM for almost everything. We have over a >> hundred model classes in in the project (~150 DB tables) and we have >> about 18 "normal" SQL queries in the app. >> >> I was curious so I just grepped the source code. Looks like we fall >> back to SQL in only a couple of cases: >> >> * Bulk import, as mentioned above >> >> * Bulk export/reporting -- although a lot of reporting is done through >> Doctrine, because it builds all the heinous joins for us :) >> >> * Pagination -- The version of Doctrine we're using has a bug that >> misreports pagination data... It pages the queries just fine, but lies >> about how many total rows exist, so we ignore what it tells us and ask >> directly via SQL. >> >> * And one other place where I'm pretty sure it should have been done >> in Doctrine. I'll chat with the developer that wrote that code, see if >> he has a good reason for it :) >> >> >> -- >> justin >> http://justinhileman.com >> >> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From justin at justinhileman.info Thu Jul 23 15:56:01 2009 From: justin at justinhileman.info (Justin Hileman) Date: Thu Jul 23 16:04:45 2009 Subject: [UPHPU] Should I Be Using an ORM? In-Reply-To: <4A684CE1.5010703@xmission.com> References: <75DEA8DC63249446B84C2133FDC75A3827290A61@VMBX115.ihostexchange.net> <4A667511.8000106@justinhileman.info> <990870EB-10AD-416F-9D8B-E5D13D885BA4@gmail.com> <4A684CE1.5010703@xmission.com> Message-ID: <4A68DC71.4040901@justinhileman.info> Walt Haas wrote: > > you would need PHP 5.2.3 or later and the PDO extension, and from > > http://www.doctrine-project.org/documentation/cookbook/1_1/en/my-first-project > > you need command line access to the server. Yep, you need PDO. That should come standard with any self-respecting shared hosting environment. You don't, however, need command line access to use Doctrine. There are PHP function calls equivalents for anything you would use the shell script to do (generate model classes, create, modify, and drop DBs, etc). The command line is easier, but it's not the only way. So lack of SSH access shouldn't be the deal breaker. -- justin http://justinhileman.com From jgiboney at gmail.com Thu Jul 23 21:11:58 2009 From: jgiboney at gmail.com (Justin Giboney) Date: Thu Jul 23 21:20:25 2009 Subject: [UPHPU] Problem with Doctrine Message-ID: I posted a question about a problem I am having with Doctrine. If any of you experts can answer it, that would be awesome. http://stackoverflow.com/questions/1175534/trouble-getting-mamp-to-work-with-pdo-mysql Thanks, Justin Giboney From justin at justinhileman.info Fri Jul 24 09:25:27 2009 From: justin at justinhileman.info (Justin Hileman) Date: Fri Jul 24 09:37:33 2009 Subject: [UPHPU] Problem with Doctrine In-Reply-To: References: Message-ID: <4A69D267.8020003@justinhileman.info> Justin Giboney wrote: > I posted a question about a problem I am having with Doctrine. If any of > you experts can answer it, that would be awesome. Part of the problem is that you're using MAMP... XAMPP >> MAMP -- justin http://justinhileman.com From joseph at josephscott.org Fri Jul 24 10:17:01 2009 From: joseph at josephscott.org (Joseph Scott) Date: Fri Jul 24 10:34:53 2009 Subject: [UPHPU] Problem with Doctrine In-Reply-To: <4A69D267.8020003@justinhileman.info> References: <4A69D267.8020003@justinhileman.info> Message-ID: <6DDD1A32-8C25-476A-8AA4-BD879D87C605@josephscott.org> On Jul 24, 2009, at 9:25 AM, Justin Hileman wrote: > Justin Giboney wrote: >> I posted a question about a problem I am having with Doctrine. If >> any of >> you experts can answer it, that would be awesome. > > Part of the problem is that you're using MAMP... XAMPP >> MAMP One of the reasons I don't use XAMPP for the Mac is that they ship with known broken combinations of PHP and libxml according to this page: http://www.apachefriends.org/en/xampp-macosx.html PHP 5.2.9 & libxml-2.7.2. They'll need to use at least version 2.7.3 of libxml (combined with PHP 5.2.9 or higher) for them to work correctly. Other wise parsing XML from PHP that uses the libxml library will result in striped brackets. -- Joseph Scott joseph@josephscott.org http://josephscott.org/ From vvilla at gmail.com Fri Jul 24 13:22:13 2009 From: vvilla at gmail.com (Victor Villa) Date: Fri Jul 24 13:30:37 2009 Subject: [UPHPU] Geek Lunch Friday July 31st @ 1pm Message-ID: Greets, So the recession is hurting and business is down, and you?re thinking, man? what to do? Here is what you do, come to Geek Lunch next Friday at Tucanos! There is one in Salt Lake and one in Utah County so it?s close to everybody. So if you just need to pull away from the office to relax or want to network with your fellow technologists, just remember this, last geek lunch at tucanos had nearly 40 attendees from all walks in the tech field. Coders, sysadmins, DBAs, local business owners, you name it, they were there. Perfect casual networking opportunity! Please please please RSVP so i can get appropriate seating at Tucanos. Here is the info: *Salt Lake County ? July 31st, 1pm* RSVP ? http://upcoming.yahoo.com/event/4111644/?ps=5 Tucanos 162 S 400 W Salt Lake City, UT 84101-1145 Get Directions (801) 456-2550 * Utah County ? July 31st 1pm *RSVP* ? *http://upcoming.yahoo.com/event/4111665/?ps=5 Tucanos 4801 N University Ave # 790 Provo, UT 84604-7745 Get Directions (801) 224-4774 From lists at sollis.net Sun Jul 26 19:22:26 2009 From: lists at sollis.net (Chad Sollis) Date: Sun Jul 26 19:30:14 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: References: <221452.53728.qm@web52011.mail.re2.yahoo.com> Message-ID: <65EEEE6F-0768-45E6-990C-37D3D9A2B486@sollis.net> Thank you for the info... any major considerations between the database options I mentioned (mysql,postgres, amazon aws, google bigtable, etc) as it relates to using logic on the db level? Thank you. On Jul 22, 2009, at 12:36 PM, Joseph Scott wrote: > > On Jul 22, 2009, at 11:32 AM, CarSign wrote: > >> allows more logic to be performed on the DB which can reduce load >> on PHP. > > > I'd suggest that this can be both a good thing and bad thing > depending on the needs of the project. > > If the project requirements are such that it has to scale to large > number of simultaneous users and page views then this would be a bad > thing. It's much easier to spread the load at the web server/PHP > level than at the database level. > > On the other side, if this app is targeted for most controlled > environments like internal business deployments, then pushing more > logic into the DB can be very helpful. You're often less concerned > about scale since you generally have an upper bound on the number of > concurrent users (number of employees). And having more logic in > the DB can make it easy for other applications to speak to the DB > without breaking things (that might be a non-issue if provide a > really good API that other apps can use). > > My current job demands using the first approach, my previous job we > did apps that used the second. It varies depending on the needs, > requirements, expectations and limits of your project. > > -- > Joseph Scott > joseph@josephscott.org > http://josephscott.org/ > > > > From lists at sollis.net Sun Jul 26 19:24:54 2009 From: lists at sollis.net (Chad Sollis) Date: Sun Jul 26 19:32:42 2009 Subject: Fwd: [UPHPU] looking for opinions/ideas References: <65EEEE6F-0768-45E6-990C-37D3D9A2B486@sollis.net> Message-ID: <6677C400-7822-4DD7-B4C5-64953D30C17D@sollis.net> Thank you for the info... any major considerations between the database options I mentioned (mysql,postgres, amazon aws, google bigtable, etc) as it relates to using logic on the db level? Thank you. On Jul 22, 2009, at 12:36 PM, Joseph Scott wrote: > > On Jul 22, 2009, at 11:32 AM, CarSign wrote: > >> allows more logic to be performed on the DB which can reduce load >> on PHP. > > > I'd suggest that this can be both a good thing and bad thing > depending on the needs of the project. > > If the project requirements are such that it has to scale to large > number of simultaneous users and page views then this would be a bad > thing. It's much easier to spread the load at the web server/PHP > level than at the database level. > > On the other side, if this app is targeted for most controlled > environments like internal business deployments, then pushing more > logic into the DB can be very helpful. You're often less concerned > about scale since you generally have an upper bound on the number of > concurrent users (number of employees). And having more logic in > the DB can make it easy for other applications to speak to the DB > without breaking things (that might be a non-issue if provide a > really good API that other apps can use). > > My current job demands using the first approach, my previous job we > did apps that used the second. It varies depending on the needs, > requirements, expectations and limits of your project. > > -- > Joseph Scott > joseph@josephscott.org > http://josephscott.org/ > > From bigdog at venticon.com Mon Jul 27 10:38:11 2009 From: bigdog at venticon.com (thebigdog) Date: Mon Jul 27 10:46:34 2009 Subject: [UPHPU] looking for opinions/ideas In-Reply-To: <65EEEE6F-0768-45E6-990C-37D3D9A2B486@sollis.net> References: <221452.53728.qm@web52011.mail.re2.yahoo.com> <65EEEE6F-0768-45E6-990C-37D3D9A2B486@sollis.net> Message-ID: <4A6DD7F3.2070101@venticon.com> > any major considerations between the database options I mentioned > (mysql,postgres, amazon aws, google bigtable, etc) as it relates to > using logic on the db level? I am always down for postgresql and I love pl/pgsql. Writing stored procs, functions and other database objects in postgresql rock. I think mysql would be another great choice as well, if you get the newer versions. I am not sure if you want to move to pgsql++ -- thebigdog From beau at open-source-staffing.com Wed Jul 29 13:50:02 2009 From: beau at open-source-staffing.com (beau@open-source-staffing.com) Date: Wed Jul 29 13:58:19 2009 Subject: [UPHPU] [JOB] PHP5 Developer =?utf-8?b?4oCTIFRlbGVjb21tdXRlIOKAkyA1MGsg?= =?utf-8?q?to_60k/year?= Message-ID: <20090729125002.8667d86e3ec984c0238a6988004ae66c.13607c466b.wbe@email.secureserver.net> My client has a small business specializing in e-commmerce software/solutions for small/medium sized businesses. They?ve spent last 3 years working on a ground up rewrite in OO PHP5, MySQL5, Ajax (Jquery), Smarty, etc. They have had live customers for over a year now but have not officially released the application yet. They are looking for someone to take over as the head developer basically. They need to be good with both "under the hood" type of stuff (lower level framework) as well as UI/AJAX, etc. They are looking for someone that has a deep ecommerce background primarily. Someone that has been involved in large scale products, not just projects. Someone that has drive, can work without direct supervision, can bring good ideas and vision, help manage others when need, etc. It would be a plus if they are savvy in FreeBSD (compiling kernels, lighttpd, MySQL, benchmarking, etc), but not a requirement. This is a 100% telecommute, full time job paying $50,000-60,000 per year. There are no benefits at present, but that might change in the future. My client is physically located in Boston, MA and if you lived nearby for occasional meetings, it would be a plus, but not required. You can reside anywhere in the USA. Candidates should have held full time positions for 2 or more years. No ?job-hoppers? please. Strong command of English a MUST since most communication will be done via telephone, email, Skype, etc. Candidates should be well-versed in OO PHP 5, MySQL, UI and AJAX and have a deep e-commerce background. When applying, please write up a paragraph or two highlighting your experience as it pertains to this job description and include your resume and salary requirements to beau-AT-open-source-staffing.com Thank you, Beau J. Gould Open Source Staffing www.open-source-staffing.com beau-AT-open-source-staffing.com From sean at exit12.org Wed Jul 29 14:42:47 2009 From: sean at exit12.org (Sean Thayne) Date: Wed Jul 29 14:51:04 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? Message-ID: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> Hey All, I've recently been doing research of whether or not to store uploading files to a database or to just save them to a file and save the filename to the db. I have read that postgres becomes very unstable with 2mb+ files but that mysql can easily handle large files. I work with a project that allows 300+ employees to upload images and documents. They need to be highly secured and they are used by multiple sites including some vmware images, so I figure that by using a mysql database to store the files it will keep them protected and keep everything tightly tied together. But I still wonder if there are any cons to storing the files in a databases? Thanks, - Sean From wadeshearer.lists at me.com Wed Jul 29 15:00:07 2009 From: wadeshearer.lists at me.com (Wade Preston Shearer) Date: Wed Jul 29 15:08:34 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> Message-ID: <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> What are the reasons that you are even considering putting them in a DB? On 29 Jul 2009, at 14:42, Sean Thayne wrote: > Hey All, > I've recently been doing research of whether or not to store > uploading files > to a database or to just save them to a file and save the filename > to the > db. I have read that postgres becomes very unstable with 2mb+ files > but that > mysql can easily handle large files. > > I work with a project that allows 300+ employees to upload images and > documents. They need to be highly secured and they are used by > multiple > sites including some vmware images, so I figure that by using a mysql > database to store the files it will keep them protected and keep > everything > tightly tied together. > > But I still wonder if there are any cons to storing the files in a > databases? > > > > Thanks, > > - Sean > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2431 bytes Desc: not available Url : http://uphpu.org/pipermail/uphpu/attachments/20090729/b8f2f344/smime.bin From thinbegin at gmail.com Wed Jul 29 15:04:02 2009 From: thinbegin at gmail.com (thin) Date: Wed Jul 29 15:12:38 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> Message-ID: <6360b3a10907291404k17711b85ha738af159dfda0dd@mail.gmail.com> Access time is the only potential issue here, and that isn't even - necessarily - much of a hit. depending on usage/needs. i've stored a *lot* of large files in mysql and been happy with the method. -thinbegin On Wed, Jul 29, 2009 at 2:42 PM, Sean Thayne wrote: > Hey All, > I've recently been doing research of whether or not to store uploading files > to a database or to just save them to a file and save the filename to the > db. I have read that postgres becomes very unstable with 2mb+ files but that > mysql can easily handle large files. > > I work with a project that allows 300+ employees to upload images and > documents. They need to be highly secured and they are used by multiple > sites including some vmware images, so I figure that by using a mysql > database to store the files it will keep them protected and keep everything > tightly tied together. > > But I still wonder if there are any cons to storing the files in a > databases? > > > > Thanks, > > - Sean > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net > From sean at exit12.org Wed Jul 29 15:05:53 2009 From: sean at exit12.org (Sean Thayne) Date: Wed Jul 29 15:14:09 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> Message-ID: <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> Because It can be a shared resource, it can handle requests on the localhost or by vmware images in the same way. I don't have to create a bunch of symlinks to folders and use nfs/samba for vmware images It is password protected I can easily attach additional information to a file, without actually touching the file I can use foreign keys to handle garbage collection of files I can easily manage additional versions of files... For instance I could cache different versions of a image. All in all I think it would just be much easier to manage a large set of files in a database, rather than a filesystem - Sean On Wed, Jul 29, 2009 at 3:00 PM, Wade Preston Shearer < wadeshearer.lists@me.com> wrote: > What are the reasons that you are even considering putting them in a DB? > > > > On 29 Jul 2009, at 14:42, Sean Thayne wrote: > > Hey All, >> I've recently been doing research of whether or not to store uploading >> files >> to a database or to just save them to a file and save the filename to the >> db. I have read that postgres becomes very unstable with 2mb+ files but >> that >> mysql can easily handle large files. >> >> I work with a project that allows 300+ employees to upload images and >> documents. They need to be highly secured and they are used by multiple >> sites including some vmware images, so I figure that by using a mysql >> database to store the files it will keep them protected and keep >> everything >> tightly tied together. >> >> But I still wonder if there are any cons to storing the files in a >> databases? >> >> >> >> Thanks, >> >> - Sean >> >> _______________________________________________ >> >> UPHPU mailing list >> UPHPU@uphpu.org >> http://uphpu.org/mailman/listinfo/uphpu >> IRC: #uphpu on irc.freenode.net >> > > From bigdog at venticon.com Wed Jul 29 15:20:22 2009 From: bigdog at venticon.com (thebigdog) Date: Wed Jul 29 15:28:40 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> Message-ID: <4A70BD16.5070202@venticon.com> > I've recently been doing research of whether or not to store uploading files > to a database or to just save them to a file and save the filename to the > db. I have read that postgres becomes very unstable with 2mb+ files but that > mysql can easily handle large files. > > I work with a project that allows 300+ employees to upload images and > documents. They need to be highly secured and they are used by multiple > sites including some vmware images, so I figure that by using a mysql > database to store the files it will keep them protected and keep everything > tightly tied together. > > But I still wonder if there are any cons to storing the files in a > databases? Basically you are allowing the database to wrap up your images. I, personally, have never seen the need to store images into the database. I tend to store locations to the images. Then you just have the webserver handle the request for the image instead of calling apache -> php -> database -> php -> apache...seems easier to just allow the client to grab the images. If you need to make modifications for the image, I usually do that 1x and then save the image off when for later use. For example, if you need to create thumbs or other images types. If you have tons of images..then you can setup the file system to deal with them (remembers the days of 24tb of images). -- thebigdog From joseph at josephscott.org Thu Jul 30 08:39:07 2009 From: joseph at josephscott.org (Joseph Scott) Date: Thu Jul 30 08:47:30 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> Message-ID: <6BC4C900-747E-4B96-A014-596071668767@josephscott.org> On Jul 29, 2009, at 3:05 PM, Sean Thayne wrote: > It can be a shared resource, it can handle requests on the localhost > or by > vmware images in the same way. > I don't have to create a bunch of symlinks to folders and use nfs/ > samba for > vmware images > It is password protected > I can easily attach additional information to a file, without actually > touching the file > I can use foreign keys to handle garbage collection of files > I can easily manage additional versions of files... For instance I > could > cache different versions of a image. > > All in all I think it would just be much easier to manage a large > set of > files in a database, rather than a filesystem Are you going to be (or planning to in the future) any sort of database replication? I haven't tested this particular case, but I'd expect the single threaded/serialized nature of MySQL replication have some fits with 20MB data fields. -- Joseph Scott joseph@josephscott.org http://josephscott.org/ From sean at skyseek.com Thu Jul 30 09:26:32 2009 From: sean at skyseek.com (Sean Thayne) Date: Thu Jul 30 09:35:17 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <6BC4C900-747E-4B96-A014-596071668767@josephscott.org> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> <6BC4C900-747E-4B96-A014-596071668767@josephscott.org> Message-ID: <69A11FD4-6669-4087-83CB-4F2469E739EB@skyseek.com> What kind of fits would you expect? The only thing with replication I would expect is that there would be a little delay because a 20mb is going to take a little longer to transfer than a 1kb text field... - Sean Thayne On Jul 30, 2009, at 8:39 AM, Joseph Scott wrote: > > > On Jul 29, 2009, at 3:05 PM, Sean Thayne wrote: > >> It can be a shared resource, it can handle requests on the >> localhost or by >> vmware images in the same way. >> I don't have to create a bunch of symlinks to folders and use nfs/ >> samba for >> vmware images >> It is password protected >> I can easily attach additional information to a file, without >> actually >> touching the file >> I can use foreign keys to handle garbage collection of files >> I can easily manage additional versions of files... For instance I >> could >> cache different versions of a image. >> >> All in all I think it would just be much easier to manage a large >> set of >> files in a database, rather than a filesystem > > > > Are you going to be (or planning to in the future) any sort of > database replication? I haven't tested this particular case, but > I'd expect the single threaded/serialized nature of MySQL > replication have some fits with 20MB data fields. > > -- > Joseph Scott > joseph@josephscott.org > http://josephscott.org/ > > > > > > _______________________________________________ > > UPHPU mailing list > UPHPU@uphpu.org > http://uphpu.org/mailman/listinfo/uphpu > IRC: #uphpu on irc.freenode.net From lists at kittypee.com Thu Jul 30 10:04:00 2009 From: lists at kittypee.com (Lonnie Olson) Date: Thu Jul 30 10:12:16 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <69A11FD4-6669-4087-83CB-4F2469E739EB@skyseek.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> <6BC4C900-747E-4B96-A014-596071668767@josephscott.org> <69A11FD4-6669-4087-83CB-4F2469E739EB@skyseek.com> Message-ID: <8bcade370907300904v2c8f409bncc4d170cdbfdb08d@mail.gmail.com> On Thu, Jul 30, 2009 at 9:26 AM, Sean Thayne wrote: > What kind of fits would you expect? > > The only thing with replication I would expect is that there would be a > little delay because a 20mb is going to take a little longer to transfer > than a 1kb text field... Yes, that is a problem. The longer binary log transfer times will introduce higher delay, so your slaves will be behind the master even more. And depending on the transfer speeds and activity of your database, it could get bad, possibly even bad enough where you will never catch up. I really want to know if storing files in a database is considered a bad idea, in regards to best practices too. I admin a database server, and my developers started down the path to storing files in the database without talking to me about it first (there are understandable reasons for this however). Anyway, here is a list of pros and cons I have experienced: Pros: * Easy to share files between web applications, including disparate servers/networks * Simpler to setup (no NFS/Samba) * One less service to administer/monitor (NFS/Samba) * Centralized data storage (keeps files with database data) Cons: * Replication can be slow * Unexpected increase in storage usage in a MySQL database volume * Backups take forever * Re-sync of a failed slave takes forever and requires more downtime on the master * mysqldump not only takes forever but will often fail on large datasets However, there are many things you can do to mitigate some these problems. I am currently in process of building the file storage infrastructure so I can encourage the devs to migrate out of the database. I don't like it. --lonnie From sean at skyseek.com Thu Jul 30 10:21:40 2009 From: sean at skyseek.com (Sean Thayne) Date: Thu Jul 30 10:30:19 2009 Subject: [UPHPU] Storing 20mb files in mysql... is it a bad idea? In-Reply-To: <8bcade370907300904v2c8f409bncc4d170cdbfdb08d@mail.gmail.com> References: <2c690d160907291342w3356cb84m817a12a602d3b601@mail.gmail.com> <790E0428-4E4A-4643-AE8C-2E2BF1511EB2@me.com> <2c690d160907291405n5ba5d788m63cde0a06c69a998@mail.gmail.com> <6BC4C900-747E-4B96-A014-596071668767@josephscott.org> <69A11FD4-6669-4087-83CB-4F2469E739EB@skyseek.com> <8bcade370907300904v2c8f409bncc4d170cdbfdb08d@mail.gmail.com> Message-ID: <5E67A440-E17A-48B3-9A12-A5DB33E94FE9@skyseek.com> Good stuff, Thanks! - Sean Thayne On Jul 30, 2009, at 10:04 AM, Lonnie Olson wrote: > On Thu, Jul 30, 2009 at 9:26 AM, Sean Thayne wrote: >> What kind of fits would you expect? >> >> The only thing with replication I would expect is that there would >> be a >> little delay because a 20mb is going to take a little longer to >> transfer >> than a 1kb text field... > > Yes, that is a problem. The longer binary log transfer times will > introduce higher delay, so your slaves will be behind the master even > more. And depending on the transfer speeds and activity of your > database, it could get bad, possibly even bad enough where you will > never catch up. > > I really want to know if storing files in a database is considered a > bad idea, in regards to best practices too. I admin a database > server, and my developers started down the path to storing files in > the database without talking to me about it first (there are > understandable reasons for this however). > > Anyway, here is a list of pros and cons I have experienced: > Pros: > * Easy to share files between web applications, including disparate > servers/networks > * Simpler to setup (no NFS/Samba) > * One less service to administer/monitor (NFS/Samba) > * Centralized data storage (keeps files with database data) > Cons: > * Replication can be slow > * Unexpected increase in storage usage in a MySQL database volume > * Backups take forever > * Re-sync of a failed slave takes forever and requires more downtime > on the master > * mysqldump not only takes forever but will often fail on large > datasets > > However, there are many things you can do to mitigate some these > problems. I am currently in process of building the file storage > infrastructure so I can encourage the devs to migrate out of the > database. I don't like it. > > --lonnie From vvilla at gmail.com Fri Jul 31 10:38:01 2009 From: vvilla at gmail.com (Victor Villa) Date: Fri Jul 31 10:46:15 2009 Subject: [UPHPU] Geek Lunch Today! Friday July 31st @ 1pm Message-ID: Greets, So the recession is hurting and business is down, and you?re thinking, man? what to do? Here is what you do, come to Geek Lunch next Friday at Tucanos! There is one in Salt Lake and one in Utah County so it?s close to everybody. So if you just need to pull away from the office to relax or want to network with your fellow technologists, just remember this, last geek lunch at tucanos had nearly 40 attendees from all walks in the tech field. Coders, sysadmins, DBAs, local business owners, you name it, they were there. Perfect casual networking opportunity! Please please please RSVP so i can get appropriate seating at Tucanos. Here is the info: *Salt Lake County ? July 31st, 1pm* RSVP ? http://upcoming.yahoo.com/event/4111644/?ps=5 Tucanos 162 S 400 W Salt Lake City, UT 84101-1145 Get Directions (801) 456-2550 * Utah County ? July 31st 1pm *RSVP* ? *http://upcoming.yahoo.com/event/4111665/?ps=5 Tucanos 4801 N University Ave # 790 Provo, UT 84604-7745 Get Directions (801) 224-4774 From lists at sollis.net Fri Jul 31 10:38:48 2009 From: lists at sollis.net (Chad Sollis) Date: Fri Jul 31 10:47:00 2009 Subject: [UPHPU] Re: looking for opinions/ideas In-Reply-To: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> References: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> Message-ID: <8c9aad0907310938u31ae64aaj26a2d73c292564c9@mail.gmail.com> In further digging, I have found Hadoop, which I assume some of you have heard about (I was kind of surprised I had never come across it before). http://hadoop.apache.org Does anyone have experience with this platform? Positive or negative? Some pretty big folks use it: http://wiki.apache.org/hadoop/PoweredBy Many thanks for your feedback. ~Chad On Wed, Jul 22, 2009 at 10:53 AM, Chad Sollis wrote: > Greetings, > > I am scoping a new enterprise grade web application. I am simply in > research mode right now, but interested in any feedback from the group as it > relates to ideas on technology, scale, cost efficiency, development, > performance, etc, in addition to creating an environment that talented > engineers could enter and find their way around very easily. > > I am planning on using PHP for a variety of reasons using AOSD/SOA style of > architecture. I am open to other suggestions if the consensus indicates > that is not the right route. > > I know this is kind of broad question, if you have experience in even one > of the areas, any feedback is appreciated. > > Ideas under consideration: > > - Frameworks / ORM (performance, scale, and customization): > - Zend > - Cake > - Codeigniter > - Doctrine - Project > - Propel > - Database (triggers, procedures, speed and scale) > - MySQL > - Postgres > - Netezza > - Amazon SimpleDB > - Architecture (scale) > - EC2 > - Mosso > - Terremark > - Managed Servers > > Many Many Thanks for your ideas and suggestions. > > ~SOL > From bigdog at venticon.com Fri Jul 31 13:12:05 2009 From: bigdog at venticon.com (thebigdog) Date: Fri Jul 31 13:20:20 2009 Subject: [UPHPU] Re: looking for opinions/ideas In-Reply-To: <8c9aad0907310938u31ae64aaj26a2d73c292564c9@mail.gmail.com> References: <8c9aad0907220953v4ed9e161p3a02c28cad9013b9@mail.gmail.com> <8c9aad0907310938u31ae64aaj26a2d73c292564c9@mail.gmail.com> Message-ID: <4A734205.3030807@venticon.com> > In further digging, I have found Hadoop, which I assume some of you have > heard about (I was kind of surprised I had never come across it before). > > http://hadoop.apache.org > > Does anyone have experience with this platform? Positive or negative? Hadoop is used for managing and processing massive amounts of data and is very efficient in doing parallel processing on that data. Think in terms of data work flows and data warehousing. It might be overkill for what you want. We looked at using it with the following items: hive, hdfs, hbase, mapreduce and pig. After we looked into in great detail we realized that we did not have the amount of data that would warrant such and enterprise system. Even though we have a ton of data into the TB on many of our databases, we came to the realization that hadoop is to much. Most relational dbs are very sufficient for very large data sets. I would recommend using one of them at the bigging and then if you need something bigger you can move to it. If you design your system with growth in mind then the migration to something bigger will not be hard. thanks, -- thebigdog