<div dir="ltr">
Hi Andrei, <br>
thanks for your interest and response.<br>
<br>
<i>>> As for the disk exclusive control and fencing mechanism
with Pacemaker,
</i><i><br>
</i><i>>> our IT vendor is proposing to use SFEX (Shared Disk
File EXclusiveness)
</i><i><br>
</i><i>>> and fence_vmware_soap (to reset the failing node via
vCenter).
</i><i><br>
</i><br>
<i>></i><i> </i><i>
What is SFEX? I could not find this abbreviation anywhere.
</i><br>
<br>
It looks like <i>"Shared Disk File EXclusiveness" </i>. Pls take
look at the following:<br>
<a class="gmail-m_5057374519026155446moz-txt-link-freetext" href="http://www.linux-ha.org/wiki/Sfex_(resource_agent)" target="_blank">http://www.linux-ha.org/wiki/Sfex_(resource_agent)</a><br>
<br>
<i>> Interesting point. In general pacemaker timeouts are
supposed to always
</i><i><br>
</i><i>> be larger than underlying infrastructure timeouts. I.e.
you need to
</i><i><br>
</i><i>> account for multipath failover as well as internal ESXi
failover. I.e.
</i><i><br>
<br>
</i>OK, in my case, two paths of FC from each of ESXi hosts to the
storage,<br>
so we consider 60+sec (30x2+some seconds) or 120sec for the timeout.
<br>
<i><br>
> if ESXi host is unresponsive long enough, it is kicked out of
HA cluster
</i><i><br>
</i><i>> and VMs are restarted elsewhere. Disk access in this
case should be
</i><i><br>
</i><i>> regulated by internal ESXXi locking.</i><br>
<br>
Suppose if the ESXi hangs longer than 120s then it comes back,<br>
fence_vmware_soap would not work timely (as the target ESXi was
hanging)<br>
and pacemaker timeout itself would not work immediately(as SW watchdog timer),<br>
then the queued I/Os before the hang would override the disks.<br>
<br>
Sorry, I can't understand how your saying "internal ESXi
locking" with SFEX<br>
would protect my concerned scenario above.<br>
<br>
>><i> So I think SFEX is valid only if combined with STONITH
IPMI for
<br>
</i>>><i> baremetal servers or even VMware hosts,
</i><br>
>><i> and we should use fence_scsi for the recent SPC-3
compliant disk storage
</i><br>
>><i> with fence_vmware_soap on VMware. Am I right?
</i><br>
<br>
<div><i>>
This depends on your storage configuration. SCSI-3 reservation
across <br></i></div><div><i>> ESXi hosts is supported only with RDM in physical compatibility
mode.</i>
</div>
<br>
Sure, my disk configuration is RDM in physical compatibility mode, as required by VMware.<br>
<br>
Thanks for further guidance or suggestions.
</div>