Yellow Bricks

by Duncan Epping

SRM and rescanning your storage twice

· ·

I just got off the phone with a former colleague. He was implementing SRM at a customer site and couldn’t get it working correctly because the VMFS volumes weren’t discovered at the recovery site. As most of you know sometimes you need to rescan your HBA’s twice before the LUNs and orVMFS volumes are discovered. When using SRM the rescan only occurs once by default. Fortunately this is a setting that can be changed in the vmware-dr.xml file:

To enable the additional rescan, edit the vmware-dr.xml file at both the protected and recovery sites to add a <hostRescanRepeatCnt> element within the <SanProvider> element. Set the value of <hostRescanRepeatCnt> to 2, as shown in the following example:
<SanProvider>
.
.
.
<hostRescanRepeatCnt>2</hostRescanRepeatCnt>
</SanProvider>

If you are doing SRM implementations it might be useful to write this one down… Especially when combined with HP EVA’s.

Reader Interactions

Comments

  1. Frank Denneman says

    So far as I know, does the second-scanning problem only occurs on Qlogic HBA’s when communicating with a HP EVA. When we swapped it with an EMULEX HBA, it resolved the LUNs in one sweep.

  2. frank says

    Guess it’s the Qlogic. Had the same issue while connected to a HP XP1024 and to a Sansymphony environment.

  3. Duncan says

    Well it’s a combination of both probably cause most implementations i’ve done included QLogic HBA and a lot of them didn’t require a second rescan.