Surebackup Replica Restore points selection

[VladV]

I have a problem with surebackup for replicas and the replicas snapshots it decides to power on to. I have not yet opened a ticket because it's Saturday and I'm not at the office but I will first thing on Monday.

I am unable to find a pattern and understand what is happening but when running a Surebackup job against our replicas, the job decides to power other previous states of the VM and not the last one. It happened with Application Groups (though adding the VMs in app group from infrastructure seems to have fixed it) but it mostly happens with linked jobs. Before running the job I have checked the Ready Replicas section and it correctly shows the last restore points.

My environment is configured as follows:

3 ESXi (esx1, esx2 and esx3) added through vCenter
1 ESX (esx3) also added as a standalone host for replication target
All source VMs are replicated as source through vCenter from esx1 and esx2 and targeted to the standalone host esx3.
The virtual lab was installed on the standalone esx3 not through vCenter

Below is a print screen that shows the current snapshot tree after running the surebackup job.

[Vitaliy S.]

Hello Vlad,

Can you please tell me what happens if you start your SureBackup job via "Start to" option? Does it also pick up not the latest snapshot?

Thank you!

[VladV]

I started the job using Start to option and the job failed. I looked through the summary and monitored through vSphere Client and I have found out that the first machine from the App Group was started without changing its NIC to the isolated network thus booting in the production network. Luckily it hasn't obtained an IP address so no harm was done. It may be a coincidence and not related to the start to option.

I don't know what's happening.

[Vitaliy S.]

Yes, snapshot selection is not related to this, anyway let's wait till you open the support case and let our technical team investigate the reason for this failure and wrong restore point selection. BTW, are all VMs affected by this behavior? If this is happening to multuple VMs, then can you please check if there is any connection between the times/restore points picked up by the job?

[VladV]

Ok, I will continue investigating with support and update this topic when I have new information.

Checking the summary for the failed surebackup job I saw that the nic was mapped

Code: Select all

9/7/2013 7:25:19 PM          Publishing
9/7/2013 7:25:13 PM          Replica point: created at 9/7/2013 5:01:28 PM, type snapshot
9/7/2013 7:25:19 PM          Replica: ELTA-DC1_replica (Status SureBackup, original location elta-vcenter\Cluster_01_HA, target location 192.168.1.103)
9/7/2013 7:25:19 PM          Virtual lab: Virtual Lab Replication
9/7/2013 7:25:19 PM          VM: ELTA-DC1 (application group, 2048 MB vRAM)
9/7/2013 7:25:19 PM          OS: Microsoft Windows Server 2008 R2 (64-bit)
9/7/2013 7:25:19 PM         [b] Network adapter 1: MAC '00:50:56:92:7F:0F', type 'vpx', network 'LAN'[/b]
9/7/2013 7:25:19 PM          Assigned roles: dns server, domain controller, global catalog
9/7/2013 7:25:19 PM          Maximum boot time: 2100 second(s)
9/7/2013 7:25:19 PM          Application initialization: 120 sec
9/7/2013 7:25:19 PM          Heartbeat test: enabled
9/7/2013 7:25:19 PM          Ping test: enabled
9/7/2013 7:25:19 PM          Script tests: enabled, 3 tests
9/7/2013 7:25:27 PM          Reconfiguring
9/7/2013 7:25:27 PM          [b]Network 1: production 'LAN', isolated 'Virtual Lab Replication LAN', mapped[/b]
9/7/2013 7:25:27 PM          Results: 1/1 network(s) mapped, 0 unmapped
9/7/2013 7:25:27 PM          Summary: 100% total pass rate
9/7/2013 7:25:46 PM          Configuring DC
9/7/2013 7:30:53 PM Error    Powering on
9/7/2013 7:25:54 PM          Waiting for OS to boot for up to 2100 seconds (stable IP algorithm)...
9/7/2013 7:25:55 PM          Note: Will proceed to the next step at 9/7/2013 8:00:55 PM or earlier
9/7/2013 7:30:53 PM Error    Error: Operation is stopped by user
9/7/2013 7:30:56 PM          Powering off
9/7/2013 7:30:58 PM          Unpublishing
9/7/2013 7:30:58 PM          Task finished

but in vSphere client the replica VM was booted with the NIC in the production lan.

I think maybe I should open two cases, one for restore point selection and one for LAN mapping.

[Vitaliy S.]

I would appreciate if you could update this topic with the resolution once everything is sorted out. Thanks!

[VladV]

Case ID: 00441017

I haven't been able to replicate the issues with wrong NIC mapping so we put that on hold until it surfaces again.

Regarding the Restore Points selection, according to support, they were able to replicate the issue and the info was passed to the senior support team. I am waiting for a conclusion on this matter to know if there will be an easy fix or should we wait for a patch.

Regards,
Vlad

[VladV]

The issue is scheduled to be fixed in patch #2. Is no one using Surebackup for replicas?

[foggy]

This is a brand new feature, so probably not many people were able to try SureReplica so far, looks like you're one of the early adopters thanks for pointing that issue out and for updating the thread with the results of your case; much appreciated!