Skip to content

25_Routine_Maintenance

Marc A. Smith edited this page Mar 2, 2017 · 1 revision

Backup Your Configuration

You should backup your ESOS configuration from time to time, or whenever your configuration changes. There are several different methods to accomplish this. You could create a "support bundle" using the TUI, and then SCP that archive (tarball) to a local computer or some other place where you can keep it safe. Using the support bundle feature avoids backing up configuration files that might contain sensitive information (eg, passwords) so this would not be a complete archive of your configuration. You can also run the support_pkg.sh command to generate the support bundle, which is equivalent to using the TUI to perform the action.

It may be better to just archive the entire running configuration, and then you can sort through any specific files if needed at a later date. To do this, simply SSH into your ESOS server, then exit to the shell (Interface -> Exit to Shell). Then at the ESOS shell, run the following command to create a tarball:

tar cvzfp /tmp/esos_conf_backup.tar.gz --exclude='rc.d' /etc /var/lib

This will produce the '/tmp/esos_conf_backup.tar.gz' which you can then scp to another machine, or storage/archive area. Another option yet may be to use your enterprise backup software (eg, Bacula or something like that) to backup the /etc and /var/lib configuration directories (maybe if they have a SSH plugin for SCP/SFTP) on a periodic basis.


Cron Jobs

A default '/etc/crontab' file exists for each ESOS server, and there are several daily and hourly jobs that run. In case you forget to sync the running ESOS configuration, a cron job runs several times a day to do this. There are also jobs for checking file systems for errors (namely the esos_* labels), rotate log files, check the system health, and synchronize the system time with an NTP server (if configured).


Log File Management

If you're debugging some daemon or service on your system (eg, the cluster stack) and you find the log files aren't being rotated fast enough, and say the root (/) file system fills, you can manually free up space by rotate the log files with the archive_logs.sh command.