View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000271 | Apache 2.x | Bug | public | 2008-08-26 10:40 | 2020-08-24 03:13 |
Reporter | emax | Assigned To | psmedley | ||
Priority | normal | Severity | crash | Reproducibility | sometimes |
Status | closed | Resolution | fixed | ||
Product Version | v2.2.1 | ||||
Summary | 0000271: Apache V2.x "could not open transfer log file" in virtual hosts | ||||
Description | [Mon Aug 18 13:51:31 2008] [error] (OS 4)OS/2 error 4: could not open transfer log file X:/logs/apache/abcde.com-access_log. Unable to open logs And after it crash/exit. Server settings (as from apache info): Server Version: Apache/2.2.9 (OS/2) PHP/5.2.6 Server Built: Jun 15 2008 13:04:47 Module Magic Number: 20051115:15 Hostname/port: mail.quasarbbs.net:80 Timeouts: connection: 80 keep-alive: 80 MPM Name: MPMT_OS2 MPM Information: Max Daemons: -1 Threaded: yes Forked: no Server Architecture: 32-bit Server Root: X:/apache Config File: X:\apache/conf/httpd.conf | ||||
Additional Information | Configuration files available only in private email, for sensible datas/privacy reasons (the servers hosts health care sensible datas among other things). | ||||
Tags | No tags attached. | ||||
|
FWIW, error 4 is #define EBADF 4 /* Bad file number */ How many virtual hosts are you creating? Does this fail if you create just a single virtual host? |
|
at the moment 28 virtual hosts, but will increase soon not tried yet only with one thanks |
|
It would appear that this is a documented issue for virtual hosts http://httpd.apache.org/docs/2.2/vhosts/fd-limits.html |
|
Steve, hope you are just joking here. My installation/configuration of apache worked perfectly (on the same server). After the migration from (reinstall) of eCS 1.1 to 1.2MR it shown the problem. I also manage some webfarms, the one of a customer of mine has an apache v2.x under ubuntu with about 500 vhosts, no problems at all with vhosts combined logs. This is a bug of the OS2-ECS build, i hope this is clear. Thanks, best regards. |
|
I add, is 28 a "large number"? I've seen installations of apache v2 with 5000 vhosts..... Steve, here we are talking about professional world, not just about an os2 hobbysts that test apache for his personal home page. OS2-ECS portings do not works since they are not tested in real production environments. The guy how face this problem (with logs) is just trying to use apache v2 for real, and now he find the same problem. Hope you too want/like eCS to have a future. For that reason i believe that applications like Apache/mysql/weasel/bind/stunnel/ftpd/injoy FW must work for real (not only for hobbyst's tests). As an IT consultant i can sell some licenses of eCS (and injoy fw) as infrastructure (firewall/proxy/url filtering/web-mail-db server) machine/solution, of course it is not suitable for use as desktop (for obvious reasons, no flash, no skype, no updated adobe reader and so on..), but on this infra. machines i'm just surviving scheduling at least 3 reboots per week and with (written by me) rexx fault daemons that restart each *now and then* all the bugged services/apps under ecs-os2 (see injoy fw, apache, squid, weasel and so on...). I strongly hope that Paul (and you) will help to have at least apache, bind and stunnel to work for "real world" environements so that i can still sell some infrastructure machines with eCS. Thanks a lot. |
|
Hi, Actually Paul's work is very reasonable and, with respect, complaints about the availability of strategic software should be sent in the direction of someone else IMHO. You also need to remember that a lot of this software requires an 'emulation' library and various problems are limitations of these libraries and/or the OS/2 equivalents of various API calls. However, I'm pretty sure that we can get all of this working. a) Have you tried Steve Levine's suggestion about trying one virtual host? Does that work? b) Is this the same server that you have reported a trap problem with on another bug report? What trap is it, what kernel are you using and what are the trap CS:EIP details please? c) If you are willing to PM me (steve@stevesharrad.co.uk) the following: 1. All OS config files (config.sys, protocol.ini, ibmlan.ini, \mptn\bin\setup.cmd, \mptn\etc\inetcfg.ini, \popopuplog.os2) and contents of your startup folder/script. 2. All application config files for the applications you are having problems with (e.g. httpd.conf, squid.conf, named.conf etc) 3. Outputs from NETSTAT -t, netstat -s, netstat -r and netstat -r sent to different text files - obtained when your services have been running. Regards, Stephen |
|
http://smedley.info/httpd.zip might be worth a try |
|
Based on a suggestion from Steven Levine, I've implemented a workaround for this (untested). Please try http://smedley.info/httpd.zip |
|
problem disappeared after i entered the line "SET SHELLHANDLESINC=60" in config.sys |
|
with the changed code you shouldn't need 'shellhandlesinc' |
|
I believe this is fixed in current code |
Date Modified | Username | Field | Change |
---|---|---|---|
2008-08-26 10:40 | emax | New Issue | |
2008-08-29 00:06 | Steven Levine | Note Added: 0000953 | |
2008-08-29 01:13 | emax | Note Added: 0000954 | |
2008-08-29 02:09 | Steven Levine | Note Added: 0000955 | |
2008-08-29 10:09 | emax | Note Added: 0000956 | |
2008-08-29 10:20 | emax | Note Added: 0000957 | |
2008-11-22 16:32 | ssha | Note Added: 0001028 | |
2008-11-29 03:41 | psmedley | Note Added: 0001046 | |
2008-11-30 04:51 | psmedley | Status | new => assigned |
2008-11-30 04:51 | psmedley | Assigned To | => psmedley |
2008-11-30 04:51 | psmedley | Note Added: 0001050 | |
2008-11-30 04:51 | psmedley | Status | assigned => feedback |
2008-12-14 21:36 | emax | Note Added: 0001078 | |
2008-12-14 22:33 | psmedley | Note Added: 0001079 | |
2011-12-31 08:28 | psmedley | Note Added: 0002106 | |
2011-12-31 08:28 | psmedley | Status | feedback => resolved |
2011-12-31 08:28 | psmedley | Resolution | open => fixed |
2020-08-24 03:13 | psmedley | Status | resolved => closed |