This article is for Zmanda Recovery Manager For MySQL (ZRM) and Amanda Enterprise (AE)
Issue Type: General
SymptomsThe Zmanda Management Console (ZMC) crashes unexpectedly. Examining the error log in /opt/zmanda/amanda/apache2/logs/error_log or /opt/zmanda/zrm/apache2/logs/error_log shows the following error string:
[notice] Apache/2.2.22 (Unix) mod_ssl/2.2.22 OpenSSL/0.9.8zg mod_fastcgi/2.4.6 configured -- resuming normal operations [notice] Apache/2.2.22 (Unix) mod_ssl/2.2.22 OpenSSL/0.9.8zg mod_fastcgi/2.4.6 configured -- resuming normal operations [notice] FastCGI: process manager initialized [warn] FastCGI: server "/etc/zmanda/.p" started [emerg] (43)Identifier removed: couldn't grab the accept mutex [emerg] (43)Identifier removed: couldn't grab the accept mutex [alert] Child returned a Fatal error... Apache is exiting! [emerg] (22)Invalid argument: couldn't release the accept mutex
Description
ZMC runs on Apache2 which uses SysV semaphores for cross-process locks and pthread mutexes for intra-process locks. If the system runs low on memory, file handlers or another resource the mutexes will not function correctly and apache will be unable to start. An OOM killer message may be seen in the message log. The mutex type can be changed by the AcceptMutex directive.
Solution:
Modify the /opt/zmanda/amanda/apache2/conf/httpd.conf file and add the following line at the end:
AcceptMutex fcntl
Then restart ZMC:
service zmc_aee restart