This article is for Amanda Enterprise (AE) v. 3.1.X
Issue Symptoms
ERROR: Unable to connect to event server.
The ZMC services are not running.
** (process:16011): WARNING **: DBI connect('database=zmc;mysql_socket=/opt/zmanda
/amanda/mysql/tmp/mysql.sock;host=localhost;user=zmc;password=8d2112fd65db3a93df641fd7e3fa31754cbc6378;
[RaiseError => 0,
PrintError => 0,
AutoCommit => 0,
ShowErrorStatement => 0,
ChopBlanks => 1,
TaintIn => 1,
TaintOut => 1]
','',...) failed: Can't connect to local MySQL server through socket '/opt/zmanda/amanda
/mysql/tmp/mysql.sock' (2) at /opt/zmanda/amanda/perl/lib/5.8.8/i86pc-solaris-thread-
multi/Logging.pm line 202
Issue Description
When starting the ZMC services ( “/etc/init.d/zmc_aee start” ), some of the ZMC services may fail to start with the error messages above.
Usually this occurs when the ZMC services did not shut down cleanly and either processes are still running, pid files were left around, or both.
Resolution
- Stop the ZMC services:
# /etc/init.d/zmc_aee stop
2. If you are running Amanda Enterprise only (and not also Zmanda Recovery Manager) please check for any “zmanda” processes that may not have been shut down:
pgrep -f zmanda -l
3. If the output from this shows any running processes and any are mysql processes, these need to be shut down cleanly to avoid corrupting the ZMC database. Please try to shut these processes down cleanly with:
/opt/zmanda/amanda/zmandactl.sh stop mysql
4. Check again if any mysql processes are still running:
pgrep -f zmanda -l
5. Once there are no mysql processes running, all others can be terminated:
pkill -f zmanda
6. Verify that no “zmanda” processes are running:
pgrep -f zmanda -l
7. With no processes running, delete any pid files that may have been left around in
/opt/zmanda/common/eventserver/tmp/: #rm /opt/zmanda/common/eventserver/tmp/*pid
8. Now try again to start ZMC services:
/etc/init.d/zmc_aee start
