So the proper way to do this would be,

xe pool-emergency-transition-to-master
xe pool-recover-slaves

xe host-list params=uuid,name-label,host-metrics-live

xe vm-reset-powerstate resident-on=UUID-OF-FAILED-MASTER --force --multiple

/opt/xensource/sm/resetvdis.py NEW-POOl-MASTER-UUID SR-UUID master

?


On Fri, Apr 26, 2013 at 3:43 PM, Jonathan Ludlam <
[email protected]> wrote:

> No, the fix is to run the scrip 'resetvdis.py' - see this bit of the
> xenserver docs:
> http://docs.vmd.citrix.com/XenServer/6.1.0/1.0/en_gb/reference.html#pool_failures
>
> This is not obvious, but has been made better in the version of xapi under
> development at the moment, so in the next release it ought to be a bit
> smoother.
>
> Jon
>
> Sent from my iPad
>
> On 26 Apr 2013, at 19:16, "[email protected]" <[email protected]> wrote:
>
> Hello,
>
> I had a master of the pool fail.  I did the following on the slave in the
> pool,
>
> xe pool-emergency-transition-to-master
> xe pool-recover-slaves
>
> xe host-list params=uuid,name-label,host-metrics-live
>
> xe vm-reset-powerstate resident-on=UUID-OF-FAILED-MASTER --force --multiple
>
> now when I try to start a VM on the old slave that is now the master, I
> get the following,
>
> Error code: SR_BACKEND_FAILURE_46
> Error parameters: , The VDI is not available [opterr=VDI
> b4354ed7-3042-4874-93b4-59a392c43027 already attached RW],
>
> I'd tried everything I can find online to fix this, the only way I'd been
> able to fix the important VMs was to vdi-forget them, then relabel/readd
> them to the VMs and start.  That works, but is that REALLY the only way to
> fix this problem?
>
> Also the VDIs were created shared, however, they are showing false under
> xe commands.
>
> Any insight?
>
> _______________________________________________
> Xen-api mailing list
> [email protected]
> http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api
>
>
_______________________________________________
Xen-api mailing list
[email protected]
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api

Reply via email to