Yesterday, I have a migration for may cacti NMS server from one Virtual Machine to other Virtual Machine. The problem is that the time error when the VM migrated.
This make the graph cannot be updated by the cacti pooler.
The time is shifted forward 3 years. If only 24 hours shift, we can just wait for the 24 hours time difference. But 3 years, its unsusable. So the solution is only delete all rrd data and crso bad, create from scratch.
This is so bad, as I don't have any backup. Last 1 year performance history data was gone.
This is bad with cacti, if only had a protection to stop the rra being updated by the pooler when time is shifting to far away.
I don't know if this can be happen with other monitoring tools like zabbix or zennos.
I like to use zabbix, but the setup to much and I have no time to try it out.
Lesson learned with CACTI :
- Backup your RRA data once every week
- Stop the pooler work when do some migration and time changing routine
- Always make sure you have backup.
0 comments:
Post a Comment