Error: Failed to recover datastore

[Richard_van_Laa]

Hello,

I,m testing SRM with the LeftHand VSA. As far as I can tell all is setup correctly. The VSAs accordingly to the free chapter of Mike Laverick book "Administering VMware Site Recovery Manager". Replication is happening and configuring the SRAs shows a replication LUN.

When I test my simple scenario I get the message: "Error: Failed to recover datastore" in step 5 (Recover Normal Priority VMs) on all 3 VMs in this scenario.

In the management GUI of the LeftHand VSAs I see that there are VolumeLists created during the test run pointing to the Remote Snapshots and Authentication Groups for those VolumeList. I see the iSCSI initiator name from the Recovery ESX server, but also the World Wide Name ID from the Fibre Channel HBAs in the same ESX server in the Auth. Groups.

The IP list in the Dynamic Discovery for the iSCSI initiator on the Recovery ESX is empty. When I put in the IP from the Recovery VSA, the test gets an error in Step 2.

i think I'm overlooking something, but I can not put my finger on it. Can somebody shine a light on this?

I'm using ESX 3.5 U1, VC 2.5 U1, SRM 1.0.0-97878, VSA 7.0.0.1

Thnx ahead, Richard

[jbloo2]

Hi Richard -

The IP address of the Lefthand VSA needs to be configured in the iSCSI initiator prior to the running the test, otherwise the initiator won't be able to log into the VSA and access the snapshot. SRM will not configure the iSCSI initiator as part of the test, it will merely rescan the storage looking for new snapshot LUNs or failed over volumes on the iSCSI target that you have configured manually. However, running the test again after configuring the dynamic target should succeed, assuming the ESX host can log into the VSA. One thing that often causes confusion is that on by default ESX sets the firewall for the iSCSI initiator to be on (even after you've configured the initiator), so that outgoing connections to storage are silently refused. You might want to make sure the iSCSI software initatior has its firewall port opened; in the VI client select the host and go to the Configuration tab, then click on Security Profile then Properties... where you should see "software iSCSI client". If the box is unchecked then it is a firewall problem. Otherwise, posting the SRM log might be helpful.

The issue with the fibre channel HBAs in the authentication list is because SRM passes all of the HBAs -- both FC and iSCSI -- on the ESX host to the adapter, and the adapter authorizes those initiators to see the volumes that SRM manages. Older versions of the Lefthand adapter did not check if the initiator was FC, but it is harmless to authorize an FC-style WWN, so this would not cause any problems. Probably you are using test ESX systems, but because Lefthand is a pure iSCSI solution normally ESX hosts don't have FC HBAs on them.

[Richard_van_Laa]

Hi jbloo2,

It works. However, I do not understand. Like I stated in my original post, I've tried the test with the IP address for the Recovery VSA in the Dynamic List of the iSCSI initiator, because it makes sense that the ESX server should be able to connect to the VSA. Otherwise, how can the snapshots LUN be added to the Storage. But this resulted in an error message in step 2.

Anyway it works now and many thanx.

Richard.