Skip to main content

Hello Community Members. 

We are using Commvault IntelliSnap with Infinidat storage. Everything works fine except the device cleanup on the Linux hosts. Most of the hosts are getting alerts for path failure and IO errors at the end of the IntelliSnap backup while the unmap and device cleanup is performed.  

 

We did check the difference between the servers which are not alerting, and which are alerting.  The main difference we can see in the IntelliSnap log is about “ISHDM IntelliSnap Linux Host Device Manager”. But I can’t get any explanation about ISHDM on Commvault site or on RedHat.

 

Can you guys help me to understand what does this value /settings does better than normal “classic OS utility interface”

 

If the above setting is not related to IO error/ path error / device cleanup, please let me know if someone experienced these types of issues on Linux hosts while performing IntelliSnap backup. 

 

Logs found on the server with no IO error / path error. 

3724588 38d52c 09/04 19:45:21 17011774 ISHDM Checking for type of OS utility interface to be used ...
3724588 38d52c 09/04 19:45:21 17011774 ISHDM IntelliSnap Linux Host Device Manager is ENABLED - mode g438 - << cleanup, includes FS, force cleanup, pre-unmap cleanup, host-wide cleanup, HDM OS utility, >>]
3724588 38d52c 09/04 19:45:21 17011774 ISHDM >>  Using ISHDM OS utility interface for FS - snap source device [/dev/mapper/adat1820d21iscsi]
3724588 38d52c 09/04 19:45:21 ######## ISHDM >>  Host device manager object empty - creating it

 

 

 

Logs found on the server with IO error / path error. 

164445 2825d 09/03 21:34:54 13517428 ISHDM Checking for type of OS utility interface to be used ...
164445 2825d 09/03 21:34:54 13517428 ISHDM IntelliSnap Linux Host Device Manager is NOT enabled v0] - using classic OS utility interface
21251 5303  09/05 03:50:39 13525577 ISHDM Checking for type of OS utility interface to be used ...
21251 5303  09/05 03:50:39 13525577 ISHDM IntelliSnap Linux Host Device Manager is NOT enabled D0] - using classic OS utility interface

 

 

 

On the nodes that are working, check to see if the “ nSnapManageHostDevices “ additional setting has been set. If so, copy the key to another nodes and use the same value as the working hosts.

 

It looks like this fix has been applied at your organization before.

 

Thank you @Damian Andre . Let me take a look.  yeah, may be you are right. someone would have applied the fix on few servers and missed to update on the one which are configured later. 

Thank you @Damian Andre . I looked at the servers after changing those settings.  

Currently we are not receiving “tur checker reports path is down” path down alerts after the change. 

But we are still getting “blk_update_request: I/O error, dev dm-93 xxxx”  errors. 

 

Do you have any other suggestions to completely remove these IO error events during the snapbackup ?

The same fix is applicable to to RHEL 7 and RHEL8 boxes ? We had one server which used to trigger lot of IO error alerts and its on RHEL7, that server is not reporting errors after we made changes. Currently all the server reports the IO errors are having RHEL 8

In the above example, the Linux Host Device Manager is enabled with mode 438.

 

Do we have any other modes available other than   438 and if yes, can we get some documentation which explain what type of operation each mode performs?

 

@Damian Andre 

Reply


Terms & ConditionsCookie settings