Getting a "Lock" Error on Decommissioned Nodes

[buster99]

Hello, Running into a very odd issue. Our TSM server us getting flooded with the following errors:
02/12/20 12:55:05 ANR0430W Session 628894 for node RDRE1-XXXXX (WinNT)
refused - node name is locked.

The nodes that are getting it exist but are decommissioned and they are not associated with a backup schedule. Why would they be trying to start sessions in TSM ? Thank you

 

[marclant]

Did you disable/uninstall the client scheduler? If not, the client scheduler will try to connect a few times a day to check if there are any schedules to run.

 

[buster99]

I doubt they were done on the client side. Would this cause a slowdown of live session processing for legit backups ? That is how we stumbled across these messages, We have nodes sitting at RecW and IdleW much longer than usual. There are quite a few nodes trying to connect. Thanks

 

[marclant]

I don’t think it would cause a slow down, the session is rejected right away. It’s certainly less stress than when they were doing backups.

RecvWait means the client was in the middle of a transaction, but stopped sending data temporarily. Slow disks on the client or networking issue. The session will get cancelled if it stays in RecvW longer than the commtimeout.

IdleWait is that the client has an opened session, but not doing any work. That’s common. Each client usually has 2 sessions by default. A producer that does determines what needs to be backed up and a consumer that does the actual data transfer. At some point during the backup, the producer gets idle and sometimes the session is cancelled when hitting the idletimeout.

 

[hogmaster]

setup the SP server OS firewall to deny connections from those clients, if this solves your issue you know where the request originate.
You can see the IP adress of the nodes in the actlog.