When restoring a SAP database (via BRTools, that is: brrecover/brrestore initiating Data Protector’s backint)
it can happen, that customers get several times errors at the step of restoring the archived logs.
The following seems to happen:
1. first backint process starts the restore session on cellserver and succeeds with restore
2. when first backint process finishes, the restore session on cellserver starts to shut down (but is still there)
3. the second backint process (or to be more precise: the spawned saprest.exe process) can still connect to the existing session (OB2_ConnectSession),
but when trying to start the actual restore
(OB2_StartObjectRestore) the session on cellserver has completed, hence the error
Data Protector cpe suggestion as a workaround in case of sap restore issue:
We are suggesting to set global variable SmWaitForNewRestoreClientSec to 300 on the cellmanager.
We have come up with this value from our observation during the testing in the lab.
We were able to reproduce the issue 1 in 10 times with default SmWaitForNewRestoreClientSec(which is 30) value.
We observed, RSM takes approximately 30 secs to shut down once it finishes with the 1st backint process.
The restore can still fail with this setting. If it fails, then this may be increased to 600 and retry the same operation.
However, there is no performance impact.
For further information please review labcase id : QCIM2A78460
Change Request QCCR2A81535 from CPE Incident.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.