Minor Description
Hi,
after the last update from Observium 19.2.9712 (18th February 2019) on last monday 18th of March the system got stuck somehow and discontinued to poll any data.
In the update.log I've found following message:
Skipped 'includes/polling/applications/ioping.inc.php' -- Node remains in conflict |
Observium was at this point in maintenance mode and was not able to be recovered automatically after the update finished.
We copied the systemfiles from another instance to the machine to get the interface run again.
After some further investigation we found out, that all devices in the database got set "status = 0".
We updated all fields manually:
UPDATE devices SET status = 1 WHERE status = 0 |
After doing this all machines got polled again, but no data was received.
It seems that something with php might be not quite right.
fping run manually from the shell has no packet loss
fping run by observium (via php) has 100% loss.
After a short test with a fping-test.php file we also had 0% packet loss if fping was run by php.
We did a rollback from an older backup so we got the monitoring again up and productive but are afraid to update again now.
An interesting sidenote: We got two instances on two virtual machines running on the same host with same specs and hardware within the same software environment.
One Observium crashed after the update, the other one didn't.
We want to try another controlled update with freshly backups within the next week.
If you got any hints or additional info what might be the cause or what we can try , please let me know.
Best regards,
Sebastian