I have 2 VM MediaAgent installed for cloud backup, both Media Agent have DDB With 4 partitions. To make a full air Gap i enable Power Manager on both Media Agent. And i created DDB verification scheduler for Cloud DDB.
But mostly of time when the scheduler start and commvault turn vm's on then some of partitions breaking down and DDB Reconstruction start.
Best answer by Mike Struening RETIRED
Sharing case solution:
update from Development to force a time wait/delay. This way there will be an extended period after the Services are requested to stop, before the VM is Force shutoff.
Apply update UpdateBundle_Build1108133_Form3802
And create the following additional settings on Commserve. Properties of the Commserve > Additional Settings.
Name : PMServiceStopWaitTimeInMins Category : CommServDB.GxGlobalParam Value : 5
The Value can be 5 to 60 ( mins ). This is the time out period of service stop request of MA. The CS will wait for SIDB process to go down till this time. If it’s not going down by this time, there will be a force power-off. If SIDB process is taking more than 5 mins to go down, then you can change this value to set an ideal time out period.
Good morning. When the VM is powered down, are Commvault services being shut down prior? If you are just powering the VM off without stopping CV services first, especially SIDB2.exe, you could see corruption of the DDB.
@Orazan Ok i can try it, but i dont want to run this scrip by the hand, because when there is no activity on MA, vm wil be power off by commvault. How can i implament the script before the commvault power off vm?
@Egor Skepko , I had a good chat with some internal folks who advised that you should turn off Data Verification on cloud libraries and see if the DDB corruption still occurs.
We don’t recommend running CV on cloud based DDBs normally.
Let me know how this goes. If it continues to corrupt, then we should open a support case.
Hi All, power management feature stops the services on MA before shutting down the VM. So there is no need to manually run any command to stop service. As part of CVD stop, it waits for DDBs to go down gracefully. However power management has a time out of 30 seconds for services to be stopped. So if the services are not stopped within 30 seconds (a bit unusual though), the machine is shut down and we may see DDB corruptions. As part of a related customer issue we are discussing this aspect. Let me know if you have any further questions.
I have disable ddb verification for cloud, but yestoday i sow again ddb reconstruction of partitions. I dunno why and what happend. The vm's were down btw. Its might happend after aux.
@Egor Skepko - the DDB corruption has nothing to do with DDB verification. As I explained above, the corruption happens when in between CVD stop the cloud VM is shut down. As said, power management waits for 30 sec before issuing a shut down, so the issue happens only when CVD does not go down in 30 sec. Is this issue escalated to support? if so can you give me the reference ticket number?
update from Development to force a time wait/delay. This way there will be an extended period after the Services are requested to stop, before the VM is Force shutoff.
Apply update UpdateBundle_Build1108133_Form3802
And create the following additional settings on Commserve. Properties of the Commserve > Additional Settings.
Name : PMServiceStopWaitTimeInMins Category : CommServDB.GxGlobalParam Value : 5
The Value can be 5 to 60 ( mins ). This is the time out period of service stop request of MA. The CS will wait for SIDB process to go down till this time. If it’s not going down by this time, there will be a force power-off. If SIDB process is taking more than 5 mins to go down, then you can change this value to set an ideal time out period.
We use 3 different kinds of cookies. You can choose which cookies you want to accept. We need basic cookies to make this site work, therefore these are the minimum you can select. Learn more about our cookies.