Submitted by clwheatley on Tue, 01/25/2011 - 14:39
I have some users trying to update a Wordpress site (plugins and CMS) using the automatic update feature but the update script constantly times out.
Increasing the Server Configuration > Website Options > Maximum PHP script run time value and restarting Apache has no effect as the script times out at 31 seconds.
I've also tried editing the PHP config and increasing both the parsing and execution times, but again, with no luck.
The site is running FCGId, though I've also tried switching to CGI wrapper.
Am I missing something?
Chris
Status:
Closed (fixed)
Comments
Submitted by JamieCameron on Tue, 01/25/2011 - 14:59 Comment #1
What error message are they getting exactly, and what gets logged to the domains logs/error_log file?
Submitted by clwheatley on Tue, 01/25/2011 - 18:24 Comment #2
The domain's error log reads:
[warn] mod_fcgid: read data timeout in 31 seconds[warn] (110)Connection timed out: mod_fcgid: ap_pass_brigade failed in handle_request function
Submitted by clwheatley on Tue, 01/25/2011 - 18:40 Comment #3
Ok, I think I figured it out. The site in question also runs in SSL. Changing the max script run time does not affect this site. Perhaps this is because I manually added the secure site because I could not install a shared cert in virtualmin to a new site (though valid on the shared cert). Something about an error that there is already a site on that port using SSL - weird...
Anyway, looking through the .conf file for this site I noticed that IPCCommTimeout is 31 for the SSL site but a different value for the standard site.
Chris
Submitted by JamieCameron on Tue, 01/25/2011 - 19:11 Comment #4
Ok, that would explain it - if you enabled SSL manually, Virtualmin won't know about the VirtualHost block on port 443, and so won't update the IPCCommTimeout directive in it. Change that manually and restart Apache, and you should be good ..
Submitted by clwheatley on Wed, 01/26/2011 - 15:38 Comment #5
Yep, that did the trick.
Thanks
Submitted by Issues on Wed, 02/09/2011 - 16:21 Comment #6
Automatically closed -- issue fixed for 2 weeks with no activity.