Back to mgt-trust

Mini How-To for on-call staff:


** Place a host in scheduled downtime

On the mgt server in the relevant DC - use the command nagios_host_sched_downtime.perl E.g.:

claus@mgt-dk:~> nagios_host_sched_downtime.perl as-var-2a

The command takes an optional second argument - the number of seconds to place a
server in downtime - the default is 900 seconds.

** Restart an individual appserver

After login on the appserver (via the relevant mgt server) use:
sudo -E /etc/init.d/startSERVICE restart terracotta - E.g.:

claus@as-var-2a:~> sudo -E /etc/init.d/startVAR-2A restart terracotta

** Make Heapdump and Restart an individual appserver

After login on the appserver (via the relevant mgt server) first start a new screen session:
'screen -S makedump /bin/bash' E.g.:

claus@as-var-2a:~> screen -S makedump /bin/bash

Then type the command makedump when you are in the screen-session:

claus@as-var-2a:~> makedump

When done just exit out normally of the screen. More on screen

** Restart a cell/cluster of appservers in rolling mode

On the mgt server in the relevant DC - use the command CL_restart
with the rolling flag - E.g.:

claus@mgt-dk:~> CL_restart vareu rolling

Use CL_restart to see which cells/clusters are available.

** Restart a cell/cluster of appservers in parallel mode

On the mgt server in the relevant DC - use the command CL_restart
with no flags - E.g.:

claus@mgt-dk:~> CL_restart vareu

Note: All appserver instances will be down during the process.
Use CL_restart to see which cells/clusters are available.

** Restart a cell/cluster of appservers incl. Terracotta

On the mgt server in the relevant DC - use the command CL_restart
with the terracotta flag - E.g.:

claus@mgt-dk:~> CL_restart vareu terracotta

Note: All appserver/terracotta instances will be down during the process.
Use CL_restart to see which cells/clusters are available

** Restart Terracotta on an appserver

After login on the appserver (via the relevant mgt server) use:
sudo /etc/init.d/startTERRACOTTA restart - E.g.:

claus@as-var-2a:~> sudo /etc/init.d/startTERRACOTTA restart

Check output on the terracotta.out logfile after using the golog command.

** Place a hostgroup in scheduled downtime

If a whole vh-platform (virtualization host - e.g. vh01-dk) and the VMs on that
vh must be taken offline - all of the VMs and the vh itself can be placed in scheduled
downtime by the use of nagios_hostgroup_sched_downtime.perl on the mgt
server in the relevant DC - E.g.:

claus@mgt-dk:~> nagios_hostgroup_sched_downtime.perl vh01-dk_guests

Notice the _guests suffix. The command takes an optional second argument -
the number of seconds to place the vh in downtime - the default is 900 seconds. If the
command does not recognized the hostgroup name - it'll present a list of known names.

Back to mgt-trust