Keep Server Online
If you find the Apache Lounge, the downloads and overall help useful, please express your satisfaction with a donation.
or
A donation makes a contribution towards the costs, the time and effort that's going in this site and building.
Thank You! Steffen
Your donations will help to keep this site alive and well, and continuing building binaries. Apache Lounge is not sponsored.
| |
|
 Topic: Problems setting up PHP 5.1.6 + eAccelerator on Win2K |
|
| Author |
|
halfer
Joined: 31 Jan 2007
Posts: 19
|
 Posted: Thu 10 May '07 11:55 Post subject: Problems setting up PHP 5.1.6 + eAccelerator on Win2K |
 |
|
Hi all
I managed to set up Apache 2.0.59 on a Windows 2000 SP4 server, with PHP 5.1.6, eAccelerator 0.95, and mod_fcgid obtained from Apache Lounge. I have two server instances, each sharing the same PHP installation, and as far as I can tell it all works perfectly.
However, on my live server (also Windows 2000 SP4 with the same PHP and eAccelerator but no fcgid yet) I am having some difficulties. I have set up two instances, and the first one seems to be working fine. However the second instance is having trouble: for some reason it fails if it encounters particular PHP exceptions, and strangely this brings down the first instance as well. I thought that instances were meant to be well insulated? I have even made a separate copy of the PHP folder, modified the relevant http.conf and php.ini files, and the same behaviour is still exhibited.
In debugging mode, the part of the second instance's error log is thus:
| Code: | [Thu May 10 10:00:47 2007] [debug] child.c(697): Child 1332: Worker thread 119 starting.
[Thu May 10 10:00:47 2007] [debug] child.c(697): Child 1332: Worker thread 121 starting.
[Thu May 10 10:00:47 2007] [debug] child.c(697): Child 1332: Worker thread 123 starting.
[Thu May 10 10:01:35 2007] [debug] mod_rewrite.c(1788): [client 127.0.0.1] mod_rewrite's internal redirect status: 0/10.
[Thu May 10 10:01:37 2007] [debug] mod_rewrite.c(1788): [client 127.0.0.1] mod_rewrite's internal redirect status: 0/10.
[Thu May 10 10:01:45 2007] [error] [client 127.0.0.1] PHP Fatal error: Uncaught exception 'PropelException' with message ' [wrapped: connect failed [Native Error: ORA-12514: TNS:listener does not currently know of service requested in connect descriptor] [User Info: Array]]' in C:\\PHP5.1.6\\PEAR\\symfony\\vendor\\propel\\Propel.php:484\nStack trace:\n#0 D:\\Inetpub\\wwwroot\\nf_framework_uat\\lib\\model\\om\\BaseUserMainPeer.php(285): Propel::getConnection('framework')\n#1 D:\\Inetpub\\wwwroot\\nf_framework_uat\\lib\\model\\om\\BaseUserMainPeer.php(266): BaseUserMainPeer::doSelectRS(Object(Criteria), NULL)\n#2 D:\\Inetpub\\wwwroot\\nf_framework_uat\\apps\\frontend\\lib\\myUser.class.php(179): BaseUserMainPeer::doSelect(Object(Criteria))\n#3 D:\\Inetpub\\wwwroot\\nf_framework_uat\\apps\\frontend\\lib\\myUser.class.php(95): myUser->lookupUserMain()\n#4 D:\\Inetpub\\wwwroot\\nf_framework_uat\\apps\\frontend\\lib\\myUser.class.php(114): myUser->setLoginDetails('_default', NULL)\n#5 D:\\Inetpub\\wwwroot\\nf_framework_uat\\apps\\frontend\\lib\\myUser.class.php(375): myUser->setDefaultLoginDetails()\n#6 C:\\PHP5.1.6\\PEAR\\symfony\\user\\sfBasi in C:\\PHP5.1.6\\PEAR\\symfony\\vendor\\propel\\Propel.php on line 484
[Thu May 10 10:02:37 2007] [notice] Parent: Received shutdown signal -- Shutting down the server.
[Thu May 10 10:02:37 2007] [notice] Child 1332: Exit event signaled. Child process is ending.
[Thu May 10 10:02:38 2007] [notice] Child 1332: Released the start mutex
[Thu May 10 10:02:38 2007] [debug] child.c(758): Child 1332: Worker thread 2 exiting.
[Thu May 10 10:02:38 2007] [debug] child.c(758): Child 1332: Worker thread 5 exiting. |
I notice that the path above "C:\\PHP5.1.6\\PEAR\\symfony\\user\\sfBasi" is not fully complete and wonder if that might have anything to do with it. Has anyone experienced behaviour like this before?
One thing that might be responsible for this behaviour is that I tried to install the Visual C++ extensions from Microsoft to enable fcgid to work. However this caused an error, which I discovered from another post here results from the Windows Installer stuff being too old. I therefore cheated and used Sysinternal's Procmon to see what DLL was missing and I copied this from the first server to the second. Should I have run these installers properly in case this is to blame?
Lastly in this second instance I have completely rewritten the httpd.conf into separate files and removed a number of Apache modules/config lines that are not in use. Might the order of statements cause problems with stability?
Thanks all in advance. |
|
| Back to top |
|
halfer
Joined: 31 Jan 2007
Posts: 19
|
| Posted: Thu 10 May '07 13:41 Post subject: Problems setting up PHP 5.1.6 + eAccelerator on Win2K (pt 2) |
|
|
I've tried something new. I renamed the httpd config file for the second troublesome instance, copied the working httpd config file into the name expected for the second instance, and found that without mod_fcgid, things worked fine. (This points to the suggestion that I had misordered things previously, in my completely rewritten httpd.conf, or perhaps I had removed too many Apache modules!)
However, if I now enable mod_fcgid in the second instance, restart it, and then try to serve a php script from it, the first instance stops listening. In fact it looks like some Apache processes have spawned and, after the second instance gives up, become zombies. Any idea why this might be, or what I can do to resolve it? |
|
| Back to top |
|
halfer
Joined: 31 Jan 2007
Posts: 19
|
| Posted: Wed 16 May '07 19:07 Post subject: |
|
|
In the absence of any replies, and given the problems I've experienced, I have abandoned getting the server in question to work with FCGI. Presently we have an automatic soft restart every two hours (and it even crashed between restarts today during the day). This is clearly not satisfactory so I am looking for a new strategy that doesn't involve porting my app to IIS.
My current line of thinking is to upgrade Apache 2.0.59 to the latest in the 2.2 branch. If I remember I shall post a note back here as to how this worked under load. |
|
| Back to top |
|
|
|
|
|
|
