I just upgraded ISPConfig from version 3.0.5.4p1 to 3.0.5.4p5 using the upgrade script, accepting all of the default options. Now Apache will not run with the error that httpd is dead but the pid file still exists. Here's my Apache error log. Code: [Tue Nov 18 02:15:22 2014] [notice] Digest: done [Tue Nov 18 02:15:23 2014] [warn] Init: Name-based SSL virtual hosts only work for clients with TLS server name indication support (RFC 4366) [Tue Nov 18 02:15:23 2014] [notice] mod_python: Creating 4 session mutexes based on 10 max processes and 0 max threads. [Tue Nov 18 02:15:23 2014] [notice] mod_python: using mutex_directory /tmp [Tue Nov 18 02:15:23 2014] [warn] [mod_spdy/0.9.4.2-413] [5064:5064:WARNING:mod_spdy.cc(166)] mod_spdy is installed, but has not been enabled in the Apache config. SPDY will not be used by this server. See http://code.google.com/p/mod-spdy/wiki/ConfigOptions for how to enable. [Tue Nov 18 02:15:23 2014] [notice] Apache/2.2.15 (Unix) DAV/2 mod_ssl/2.2.15 OpenSSL/1.0.1e-fips mod_fcgid/2.3.9 mod_python/3.3.1 Python/2.6.6 mod_ruby/1.3.0 Ruby/1.8.7(2011-06-30) configured -- resuming normal operations [Tue Nov 18 02:16:01 2014] [emerg] (43)Identifier removed: couldn't grab the accept mutex [Tue Nov 18 02:16:01 2014] [emerg] (43)Identifier removed: couldn't grab the accept mutex [Tue Nov 18 02:16:02 2014] [alert] Child 5068 returned a Fatal error... Apache is exiting! [Tue Nov 18 02:16:13 2014] [emerg] (22)Invalid argument: couldn't release the accept mutex
I was able to get around the issue but adding this to my http.conf file. Code: AcceptMutex fcntl But the question still remains, why is this happening after running the ISPC update and how do I fix this properly?
I did not had that problem yet and it might be that the issue was triggered by the apache restart that the update did and not by the update itself (as it does not change the apache config except of the ispconfig vhost). But as far as I can see, the solution you aplied is the corect one for that error.
Ok, there are a whole lot of other problems too. I just noticed that all mail is now being deferred. I can't get any of it to go through. Also, I'm being blasted by thousands of emails saying "WARNING - Clearing semaphores table for user apache." I think Apache is crashing and reloading on every page request. I'm using the Perfect server setup on CentOS 6.5.
I used the backup option to /var/backup. Is it possible to reverse the upgrade and restore my last configuration from that?
Sure, thats possible, but I dont think that the ispconfig update is the reason for that problem. To restore it, unpack the backup which is in a subfolder of /var/backup/ and move the files back to their original locations. The backup contains a tar.gz file for /etc, one for /usr/local/ispconfig and a sql dump for the dbispconfig database which can be restored with phpmyadmin.
Apache seems to have start working again through no effort of my own. I left it alone overnight. Postfix still is not delivering any mails. I believe amavis may be to blame. amavisd is listening on port 10024, but I am not able to telnet on that port. I get "connect failed". Restarting related services has no effect. Also, I would really like to disable the httpd rescue emails. I am getting "httpd is down" emails by the thousands. It makes it very hard to determine what the mail problem is when I have so many of them backlogged.
Well, I've narrowed down the mail issue to amavis. When I comment out the following line in postfix's main.cf, everything begins working again. Code: content_filter = amavis:[127.0.0.1]:10024 The problem is that *should* work. Nothing about my configuration has changed. I have verified that amavis is in fact listening on port 10024 and that it is on the 127.0.0.1 interface. I'm also able to telnet on that port and interface and I do get the amavis-new greeting. I just don't know what else to try. I was also able to stop the rescue emails by editing the lib file and having it check the correct interface. My server does not have Apache listening on "localhost".
This is a VPS. I have not tried reinstalling amavis. I was working fine and I can't find where anything changed. Very strange. I've even scoured the error logs with no luck. I'll try reinstalling it.
The amavisd config has not been changed for several versions, so that it has not changed is the result that I have exected. So the ispconfig update has likely only been the trigger to show the issues due to the restart of services and not the cause. My guess is that a limit of your vps has been reached as these rather unspecifoc symptoms in many services are typical for such a limit hit in a vps. Please post the output of: cat /proc/user_bancounters
Sorry, I just caught your typo. Code: Version: 2.5 uid resource held maxheld barrier limit failcnt 474: kmemsize 166178487 548478976 1073741824 1073741824 0 lockedpages 0 11 262144 262144 0 privvmpages 829569 3369390 9223372036854775807 9223372036854775807 0 shmpages 14213 16165 9223372036854775807 9223372036854775807 0 dummy 0 0 9223372036854775807 9223372036854775807 0 numproc 182 222 9223372036854775807 9223372036854775807 0 physpages 516629 524599 0 524288 101 vmguarpages 0 0 9223372036854775807 9223372036854775807 0 oomguarpages 431913 538597 9223372036854775807 9223372036854775807 33 numtcpsock 64 146 9223372036854775807 9223372036854775807 0 numflock 17 48 9223372036854775807 9223372036854775807 0 numpty 1 2 9223372036854775807 9223372036854775807 0 numsiginfo 0 33 9223372036854775807 9223372036854775807 0 tcpsndbuf 1273648 8635904 9223372036854775807 9223372036854775807 0 tcprcvbuf 1048576 2392064 9223372036854775807 9223372036854775807 0 othersockbuf 331608 1040672 9223372036854775807 9223372036854775807 0 dgramrcvbuf 0 13080 9223372036854775807 9223372036854775807 0 numothersock 241 307 9223372036854775807 9223372036854775807 0 dcachesize 107296346 498910936 536870912 536870912 0 numfile 5099 5755 9223372036854775807 9223372036854775807 0 dummy 0 0 9223372036854775807 9223372036854775807 0 dummy 0 0 9223372036854775807 9223372036854775807 0 dummy 0 0 9223372036854775807 9223372036854775807 0 numiptent 37 40 9223372036854775807 9223372036854775807 0
Yes, sorry for the typo As you can see in the last column, your server hit 2 limits in the vm. I guess the pyspages is the likely candidate. Here a short explanation what physpages is: https://openvz.org/VSwap Is this your own server or did you rent the vm so you dont have access to the virtual machine parameters?
