...
- Write script to delete old log files if disk is at 95%
- Set up stunnel from compute nodes to controller for SMTP
Set up stunnel from controller to compute nodes for rsync to grab glance images?(unnecessary now that glance and nova will have a GlusterFS backend)- Write script to monitor nova, cinder services and notify/restart if necessary
- Write scripts to monitor resource usage
- Install larger disk on controller?
- Set up syslog aggregator on controller
- Forward IPMI syslog to controller
- Forward node syslog to controller
- Move /var/spool/mqueue on controller to a new partition; when a service goes haywire and logs fill the disk, we can't receive email warnings
- Move database dumps from ~root/backups to a different space (perhaps GlusterFS?)
- Send compute3 in for repair
- Move /var on controller to the larger /home partition; should buy time if logs are filling (20x more space there)
- Write scripts to monitor actual disk space for gluster volumes
- Put scripts into repository and check out to /usr/local/scripts on nebula
- Create option on node installation script to not run the global commands (if they've already been run, for example)