SAP_REORG_ORPHANED_TEMSE_FILES job causing application server to lose access to sapmnt share

[utham_kumar]

Hello ,
We have noticed a situation where application server loses access to sapmnt share intermittently. Whenever this issue happens it is always the application server on which SAP_REORG_ORPHANED_TEMSE_FILES job is active.

Other application servers of the same system are not impacted

This job somehow causes the application server to lose access to sapmnt share and, as such, the SAP application on the host hangs. Are you aware of any issue with this job that can trigger sapmnt access issue?

This job runs once every week but the issue only happens intermittently

It is impacting multiple Prod systems (ECC, BW - other systems in our landscape) again not always but intermittently and a common factor is this job

We are running kernel 753 PL 921 and SAP_BASIS 750 SP18 in Fiori Prod . Applications servers are running on Windows Server 2016 and DB is HANA 2.0 SPS05 Rev 55 on SLES12 ..

JOBLG folder in Fiori prod only has around 65K files - so it's not a lot

BW prod has 250 K and ECC prod has 800K files

Wanted to check if any customer has seen this similar issue and the resolution that they have applied for the same?

[utham_kumar]

Hi balaji.srinivassingh ,

Thanks for the details. In our landscape , ASCS and PAS are in different servers, and the DIR_GLOBAL is mounted in ASCS server and its accessible across app servers. We have worked with windows server team and came to know about there is a bug in SMB and it requires OS patching. However this is occurring frequently only in SAP BW system. As of now we are canceling the job and restarting the app server as a workaround to overcome this issue.

Thanks,

Uthamkumar

[bglobee]

Hi utham.kumar

as a workaround, schedule the temse cleanup job to run on CI/PAS instance only (where DIR_GLOBAL is originally located). It might be a case that App servers are facing intermittent Network connectivity issue to the SMBshare (sapmnt one).

If you want to analyze the root cause further, then setup the SMB monitoring for sapmnt at windows level , check if any backups or VM snapshots are run during the temse cleanup job run, check for any network errors/events during this time. Unfortunately, there is no mention about this error in SAP notes (no pointing to a precise root cause). So for now workaround would do.