Hi Fred,

 

  thank you for the explanation. I restarted VDSM and will monitor the behaviour.

 

Does that faulty multipath report have any side effects on stability and performance?

 

All the best,

Oliver

 

Von: Fred Rolland [mailto:frolland@redhat.com]
Gesendet: Sonntag, 11. März 2018 11:23
An: Bruckner, Simone <simone.bruckner@fabasoft.com>
Cc: users@ovirt.org
Betreff: Re: [ovirt-users] Faulty multipath only cleared with VDSM restart

 

Hi Simone,

The multipath health is built on VDSM start from the current multipath state, and after that it is maintained based on events sent by udev.

You can read about the implementation details in [1].

It seems that in your scenario, either udev did not sent the needed clearing events or that Vdsm mishandled them.

Therefore only restart of the Vdsm will clear the report.

In order to be able to debug the issue, we will need Vdsm logs with debug level (on storage log) when the issue is happening.

 

Thanks,

Fred

[1] https://ovirt.org/develop/release-management/features/storage/multipath-events/

 

On Fri, Mar 9, 2018 at 1:07 PM, Bruckner, Simone <simone.bruckner@fabasoft.com> wrote:

Hi,

 

  after rebooting SAN switches we see faulty multipath entries in VDSM.

 

Running vdsm-client Host getStats shows multipathHealth entries

 

"multipathHealth": {

  "3600601603cc04500a2f9cd597080db0e": {

    "valid_paths": 2,

    "failed_paths": [

      "sdcl",

      "sdde"

    ]

  },

  …

 

Running multipath –ll does not show any errors.

 

After restarting VSDM, the multipathHealth entires from vdsm-client are empty again.

 

Is the a way to clear those multipathHealth entires without restarting VDSM?

 

Thank you and all the best,

Simone

 

_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users