View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000299 | Apache 2.x | Bug | public | 2008-12-11 00:30 | 2020-08-24 03:12 |
Reporter | ssha | Assigned To | psmedley | ||
Priority | normal | Severity | major | Reproducibility | always |
Status | closed | Resolution | duplicate | ||
Product Version | v2.2.1 | ||||
Summary | 0000299: Apache + PHP5 crashes regularly under light load | ||||
Description | When using Apache and PHP5 to run Jinzora, Apache sometimes (and quite regularly) crashes. I have not been able to reproduce an exact request or event that will cause Apache to crash - sometimes it will crash in the night all of its own accord. The most frequent message is: [Thu Dec 04 13:21:33 2008] [error] caught exception (XCPT_ACCESS_VIOLATION) in worker thread, initiating child shutdown pid=175 [Wed Dec 10 22:25:29 2008] [error] caught exception in worker thread, initiating child shutdown pid=342 apr_canonical_error: Unknown OS/2 error code 10004 [Wed Dec 10 22:42:41 2008] [error] (OS 10004)Unknown error -1: apr_socket_accept [Wed Dec 10 22:42:41 2008] [error] (OS 10004)Unknown error -1: apr_socket_accept [Wed Dec 10 22:42:41 2008] [error] (OS 10004)Unknown error -1: apr_socket_accept [Wed Dec 10 22:42:41 2008] [notice] caught SIGTERM, shutting down (The events at 22:42:41 are caused by my CTRL-C the dead Apache window to close it). Sometimes tho an exception is raised in PHP: Killed by SIGSEGV pid=0x00af ppid=0x00ad tid=0x0001 slot=0x00b6 pri=0x0200 mc=0x0001 C:\APACHE2\BIN\HTTPD.EXE PHP5 0:001ef9b1 cs:eip=005b:1d57f9b1 ss:esp=0053:0012fe20 ebp=0012fe48 ds=0053 es=0053 fs=150b gs=0000 efl=00210246 eax=00000000 ebx=00504368 ecx=00000000 edx=00000000 edi=00000000 esi=00000000 Process dumping was disabled, use DUMPPROC / PROCDUMP to enable it. Still other times an entry is logged into POPULOG.OS2 (popuplog entries are not raised for the XCPT_ACCESS_VIOLATION error): ------------------------------------------------------------ 12-10-2008 21:07:34 SYS3175 PID 0149 TID 0003 Slot 00c8 C:\APACHE2\BIN\HTTPD.EXE c0000005 1ffd0844 P1=00000008 P2=00000258 P3=XXXXXXXX P4=XXXXXXXX EAX=00000000 EBX=0056126a ECX=02590149 EDX=17329827 ESI=0259ffa7 EDI=0259fee0 DS=9807 DSACC=00f3 DSLIM=0000117f ES=980f ESACC=00f3 ESLIM=00001e5d FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=ffef:00000844 CSACC=00df CSLIM=0000fdf8 SS:ESP=12cf:0000fe04 SSACC=00f3 SSLIM=0000ffff EBP=02590000 FLG=00010246 DOSCALL1.DLL 0003:00000844 ------------------------------------------------------------ 12-10-2008 23:16:33 SYS3171 PID 023c TID 0003 Slot 00ca C:\APACHE2\BIN\HTTPD.EXE c0000005 1d6196bf P1=00000002 P2=0257fffc P3=XXXXXXXX P4=XXXXXXXX EAX=02580020 EBX=02585a68 ECX=00000000 EDX=00000008 ESI=00763160 EDI=007a67a0 DS=0053 DSACC=f0f3 DSLIM=ffffffff ES=0053 ESACC=f0f3 ESLIM=ffffffff FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=005b:1d6196bf CSACC=f0df CSLIM=ffffffff SS:ESP=0053:0257fffc SSACC=f0f3 SSLIM=ffffffff EBP=025806b4 FLG=00010202 PHP5.DLL 0001:002896bf Remote access to this system can be arranged, if required. Because of the regularity of this problem, I suspect that there is some kind of generic fault with Apache - certainly the XCPT_ACCESS_VIOLATION when the system is not busy at all seems very strange). I am also working on the premise that no code error in the Jinzora PHP code should be able to cause this - I can't find others having this problem, at any rate. | ||||
Additional Information | To reproduce this problem with Jinzora: 1. Install Paul's PHP/MySQL/Apache 2. Download Jinzora 2.7.5 stable and unzip to c:/jinzora2 3. Set an alias in httpd.conf for Jinzora2 4. Setup Jinzora and import a little media (not much is required) 5. Use Jinzora A crash will soon happen :( Hope this you can help! | ||||
Tags | No tags attached. | ||||
|
Hi Paul, Hope you are OK. Still happens with latest Apache + PHP (2.2.11 + 5.2.8) Also now see the following error: LIBC PANIC!! fmutex deadlock: Owner died! 0x0050013c: Owner=0x814e0009 Self=0x814e0001 fs=0x3 flags=0x0 hev=0x0001004f Desc="LIBC Heap" pid=0x814e ppid=0x814c tid=0x0001 slot=0x00f8 pri=0x0200 mc=0x0000 C:\APACHE2\BIN\HTTPD.EXE Process dumping was disabled, use DUMPPROC / PROCDUMP to enable it. As this clearly shows HTTPD having at least one problem (whether or not it is the same as above), would you like me to open a seperate bug report for Apache? The above error happens when you ask a big PHP task - e.g. a re-scan of media under Jinzora. Thanks, Stephen |
|
I will try and setup Jinzora here - no need for a separate ticket - it's all related to Jinzora :) |
|
Hi Paul, Thanks :) Am using Jinzora 2.75 - JZ275.zip from the website. Remote access to my NAS/Media box is available, if you want - via VNC. PM me to steve@stevesharrad.co.uk if you need access. Thanks again, Stephen |
|
http://smedley.info/httpd.zip tidies up a few things and _may_ help |
|
http://smedley.info/php5.zip tidies up a few PHP related issues and _may_ also help |
|
Still fails now and then, even with both updates: 02-27-2009 21:41:28 SYS3171 PID a40c TID 0008 Slot 00a3 C:\APACHE2\BIN\HTTPD.EXE c0000005 1d66df5c P1=00000002 P2=0262ffd4 P3=XXXXXXXX P4=XXXXXXXX EAX=02a347e8 EBX=00000014 ECX=00d0c240 EDX=00a66ca0 ESI=009d2048 EDI=02630698 DS=0053 DSACC=f0f3 DSLIM=ffffffff ES=0053 ESACC=f0f3 ESLIM=ffffffff FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=005b:1d66df5c CSACC=f0df CSLIM=ffffffff SS:ESP=0053:0262ff9c SSACC=f0f3 SSLIM=ffffffff EBP=02630024 FLG=00010216 PHP5.DLL 0001:0031df5c ------------------------------------------------------------ 02-27-2009 21:41:52 SYS3171 PID 8ec9 TID 0008 Slot 0100 C:\APACHE2\BIN\HTTPD.EXE c0000005 1d66df5c P1=00000002 P2=0262ffd4 P3=XXXXXXXX P4=XXXXXXXX EAX=01309c80 EBX=00000014 ECX=01a2b2a0 EDX=00fa9d20 ESI=00cc3dd8 EDI=02630698 DS=0053 DSACC=f0f3 DSLIM=ffffffff ES=0053 ESACC=f0f3 ESLIM=ffffffff FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=005b:1d66df5c CSACC=f0df CSLIM=ffffffff SS:ESP=0053:0262ff9c SSACC=f0f3 SSLIM=ffffffff EBP=02630024 FLG=00010216 PHP5.DLL 0001:0031df5c ------------------------------------------------------------ 02-27-2009 22:52:58 SYS3175 PID c3c7 TID 0003 Slot 0124 C:\APACHE2\BIN\HTTPD.EXE c0000005 1ffd0844 P1=00000008 P2=00000164 P3=XXXXXXXX P4=XXXXXXXX EAX=00000000 EBX=1ffca330 ECX=0259fef4 EDX=17349827 ESI=0259fef4 EDI=00000000 DS=0053 DSACC=f0f3 DSLIM=ffffffff ES=0053 ESACC=f0f3 ESLIM=ffffffff FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=005b:1ffc9f77 CSACC=f0df CSLIM=ffffffff SS:ESP=0053:0259fb34 SSACC=f0f3 SSLIM=ffffffff EBP=0259fb7c FLG=00000246 DOSCALL1.DLL 0003:00000844 Apache error log shows: [Fri Feb 27 22:51:07 2009] [error] caught exception (XCPT_ACCESS_VIOLATION) in worker thread, initiating child shutdown pid=50119 [Fri Feb 27 22:51:07 2009] [error] caught exception in worker thread, initiating child shutdown pid=50119 [Fri Feb 27 22:53:08 2009] [notice] caught SIGTERM, shutting down With this failure, the Apache process is still running - but not serving anything :( |
|
And again... Apache seems to become exhausted, i.e. causing a Jinzora re-scan causes thousands of requests as each album is shown in the scan window... ------------------------------------------------------------ 03-13-2009 22:37:38 SYS3175 PID 501a TID 0004 Slot 0115 C:\APACHE2\BIN\HTTPD.EXE c0000005 1ffd0844 P1=00000008 P2=00000164 P3=XXXXXXXX P4=XXXXXXXX EAX=00000000 EBX=1ffca330 ECX=025bfef4 EDX=17349827 ESI=025bfef4 EDI=00000000 DS=0053 DSACC=f0f3 DSLIM=ffffffff ES=0053 ESACC=f0f3 ESLIM=ffffffff FS=150b FSACC=00f3 FSLIM=00000030 GS=0000 GSACC=**** GSLIM=******** CS:EIP=005b:1ffc9f77 CSACC=f0df CSLIM=ffffffff SS:ESP=0053:025bfb34 SSACC=f0f3 SSLIM=ffffffff EBP=025bfb7c FLG=00000246 DOSCALL1.DLL 0003:00000844 |
|
You could try increasing the number of file handles available for each process. SET SHELLHANDLESINC=50 then start httpd |
|
Look for multiple copies of php5.dll on your drive(s). Make sure Apache is loading the one that matches modphp5.dll. |
|
I'm curious if http://smedley.info/httpd-2.2.11-os2-20090728.zip helps in any way |
|
See bug 0000355 for some feedback on testing - a bit better, but memory usage seems to be the key. |
|
Please try http://download.smedley.info/httpd-2.2.15-os2-20100711.zip |
|
httpd 2.2.16, mysql 5.1.49, php 5.2.14 I have a similar problem. I suspect the actual culprit is mysql. I finally updated AMP last week. All seemed to be well until I attempted to connect to a remote mysql database, i.e., not one on the localhost. Then httpd quietly quits. No messages, nothing in the logs, poof!. It's pretty reliable about quitting. Although I did have it stop once while accessing a localhost database, but after restart it worked so it's a bit erratic for the localhost. |
Date Modified | Username | Field | Change |
---|---|---|---|
2008-12-11 00:30 | ssha | New Issue | |
2008-12-12 03:28 | psmedley | Status | new => assigned |
2008-12-12 03:28 | psmedley | Assigned To | => psmedley |
2009-01-01 15:22 | ssha | Note Added: 0001087 | |
2009-01-02 07:17 | psmedley | Note Added: 0001092 | |
2009-01-04 14:47 | ssha | Note Added: 0001093 | |
2009-02-05 09:42 | psmedley | Note Added: 0001110 | |
2009-02-22 09:18 | psmedley | Note Added: 0001149 | |
2009-02-22 09:18 | psmedley | Status | assigned => feedback |
2009-02-27 23:43 | ssha | Note Added: 0001221 | |
2009-03-14 14:57 | ssha | Note Added: 0001249 | |
2009-03-14 22:31 | psmedley | Note Added: 0001255 | |
2009-03-15 01:16 | David McKenna | Note Added: 0001256 | |
2009-07-30 21:48 | psmedley | Note Added: 0001383 | |
2009-08-08 15:49 | ssha | Note Added: 0001390 | |
2010-07-11 03:12 | psmedley | Note Added: 0001638 | |
2010-08-12 07:35 | jimoe | Note Added: 0001702 | |
2011-12-31 08:27 | psmedley | Relationship added | duplicate of 0000425 |
2011-12-31 08:27 | psmedley | Status | feedback => resolved |
2011-12-31 08:27 | psmedley | Resolution | open => duplicate |
2020-08-24 03:12 | psmedley | Status | resolved => closed |