Hi Everyone,
We currently have a vcenter estate whereby someone has thin provisioned all VM's (about 30) and overcommitment is possible on one of the LUN's.
I have set up an alarm to monitor the usage when it reaches 80% (Yellow) and 90% (Red). In normal day-to-day activities the usage will idle at about 86%.
Every so often I will receive an alert to say the datastore usage will rise to 99% for about 15-16 minutes. and then it will return to 86%. There's no additional VM's or file transfers on these LUNs and the Tasks and Events tab doesn't point to what's causing it. The additional space is approximately 146GB. No backups are being created and no snapshots are being created. The last two times this has happened it's been at different times in the day, and on different days.
I'm wondering if this could be a false positive. Prior to this warning I see event messages such as:
Virtual machine XXXX is connected
Virtual machine XXXX is connected
Virtual machine XXXX is connected
XXXXX on host is disconnected
XXXXX on host is disconnected
XXXXX on host is disconnected
Cannot synchronise host XXXX. Operation timed out.
Successfully restored access to LUN XXX
My question is : can network dropouts (iSCSI in this case) cause false positives on in these alarms?
Cheers
David