veeam 7 3par storage snapshot backups

[kte]

How is this working? Must I put all the VM's of 1 lun in the same backup job?
Or will it take a snapshot of all the data stores , of each VM a snapshot of that lun ??

K

[Gostev]

Hi Koen,

Regarding how it works, have you had a chance to review the corresponding chapter in the User Guide?

The only requirement is that all disks for the particular VM must be located on the 3PAR LUNs. The job will only use storage snapshots for VMs meeting this requirement. Other than that, there are no special considerations (for example, no need to create a separate backup job, group VMs from the same LUN into the same job, or anything like that).

Thanks!

[kte]

i check the user guide but it is no so clear.
If I have 8 4TB datastores and 4 jobs with 4 parrallel processing will veeam anlyse all the vm's and snapsht all luns on the same time on the 3 Par

K

[Gostev]

Each LUN will be snapped as required for processing the specific VM. We do not snap all LUNs at the same time.
However, if more than one VM in the job is located on the given LUN, we will process all such VMs with a single storage snapshot.

[kte]

so the jobs checks all the vm in that job on the first lun then it thakes a snapshot.
so if i have a parrallel processing of 8 with each vm on a different datastore it thakes 8 storage snapshots

K

[foggy]

Correct.

[kte]

does he present the snapshots to the backup servers also?? and rescan the veeam storage so it can do a san backup ?

K

[Vitaliy S.]

The HP SAN volume snapshot gets mounted as a new volume to the backup proxy, so yes you should have direct FC connection to your HP 3PAR storage from the backup proxy server to backup VMs in direct SAN mode.

[veremin]

and rescan the veeam storage so it can do a san backup ?

In fact, it’s not about rescanning Veeam storage/repository, but rather about rescanning VMFS on HP SAN volumes. And this depends on type of HP storage. While backup proxy is required for rescan of HP StoreServ storage system, it’s not required for rescan of HP StoreVirtual SAN storage. Thanks.

[kte]

it doesn't seem to work the mounting to the backup server, he doesn't find the lun with a certain WWN

[veremin]

For detailed information regarding preliminary steps that should be taken prior to using storage snapshots, kindly, refer to the User Guide (p.380).

If you’ve followed all the steps mentioned there and still experience certain issues, please, open a ticket with our support team and let them investigate it directly, since it’s rather hard to make any speculations without seeing actual environment or corresponding logs.

Thanks.

[kte]

case number is 00434720.

[veremin]

Thank you. Let’s wait and see what support team says about this issue.

[kte]

support called and is investigating, it seems like the veeam server doesn't find the correct luns

[cjack03]

I have a similar issue using snapshots on a 3PAR array. Most of the time the backup works smoothly and ohter times I have this same issue of the not finding the LUN with a certain WWN.

Support case is 00432752. I thought it was just me, perhaps there is something else...

[kte]

For the moment the management service of the 3par is craching alot, so I can't us the veeam integration, HP is working on it

K

[kte]

updated the call again, it is still not working he doesn't find the lun

[Brad.Sims]

cjack03,

I have also had the same error with 3par storage snapshots... "Could not find volume with Node WWN: blah blah and LUN ID #"
I opened a support case as well and have not found a fix yet.

[kte]

cjack03 , how did you configure this are all you're datastores presented to the backup sever or does veeam present only the snapshots

k

[veremin]

Also, once the investigation is finished and storage snapshots start working as expected, don’t forget to download recently released hotfix that is supposed to address the newly found issue with this functionality. Thanks.

[chjones]

Hi all,

Did you have any luck with your support cases with this issue? I am testing integration with our new HP 3PAR 7400 before we migrate live VMs onto the system and I receive the same error "Error: Unable to find volume with Node WWN xxxxxxxxxx and LUN ID x" (replaced the real values with x).

I can backup the same test VMs fine using SAN if I dont enable Storage Snapshots in the Backup Job.

[veremin]

Hi, Chris.The previous cases are still under investigation, so, it stands to reason to open your own one and let support team collect as much information regarding this issue as possible. Thanks.

[vixtech]

Hi All,

I also currently have a tech support case open for this issue. In my investigations the original WWN error referred to the FCoE capability of an Intel x520 converged 10GbE adapter installed in our physical proxy. We are not using this for SAN access as we have a dedicated Q-Logic fibre-channel host-bus-adapter handling this. This caused Veeam to pass the 3PAR incorrect information as it enumerates the cards in the machine. I have now disabled the FCoE functionality which has got rid of this particular error. Unfortunately this has not fully resolved the issue as Veeam now reports the same error for a different WWN - plugging this into a MAC/WWN lookup this WWN is manufactured by 3pardata. I have had some success with snapshot storage backup jobs but still experience intermittent "Error: Unable to find volume with Node WWN <xxxxxxxxxx> and LUN ID <x>" so cannot rely on these jobs. When it works, it's very quick.

[Gostev]

Unfortunately, we cannot test all possible deployment scenarios in our lab, so some issues will have to be caught "in the wild". Please, be sure to open a support cases, as the debug logs help our developers tremendously. And if you do this now, the fix will still make it into the next patch.

[chjones]

Thanks for the reply, Vladimir. We've opened a support case and have been submitting logs and screenshots.

What vixtech says raised an interesting thought for me. Our ESX Cluster is made up of a number of HP ProLiant BL460c Gen8 Servers that reside in multiple HP c7000 Blade Enclosures (we do this for redundancy if we lose an entire enclosure). In each enclosure we have HP FlexFabric Virtual Connect Modules which are used for both 10Gb Ethernet and 8Gb Fibre Channel to our blades. These modules allow Direct-Attach to the 3PAR (you can connect fibre directly from the Enclosure to the 3PAR, no need for Fibre Switches ... its so much less complicated and very easy to setup). This is how we have our newly installed 3PAR arrays configured.

In order for the Blade Server to access the 3PAR it must have a FlexFabric LOM, which our Gen8 Blades used for the ESX Cluster have.

However, our Veeam Proxies are older G1 blades which do not support FlexFabric. These blades have older 4Gb Fibre Channel Cards in them and connect via older Fibre Channel Modules in the Blade Enclosures to our Brocade Fibre Switches. The 3PAR, whilst cabled directly to the enclosures to use Direct-Attach, is also connected to our Brocade Switches. The 3PAR can been accessed via both Direct-Attach and via the SAN Fabric.

The Veeam Proxies access the 3PAR via the Fabric, and the ESX Hosts access via Direct-Attach.

If I do not enable Storage Snapshot Integration there are no issues backing up VMs via the SAN method. That shows me that the differences in how the ESX hosts and Proxies access the 3PAR shouldn't case any problems as Veeam has no issues backing up directly via the SAN Fabric as long as storage snapshots are not used.

I don't think this setup is the cause, but thought I'd throw it out there.

[dellock6]

Chris, have you tried the opposite, that is using one of the new blades as physical Veeam proxies? Obviously if you have at least one blade available... Not sure about this would be the cause, but just to see if anything changes...

Luca.

[Tobias_Elfstrom]

Just thought I'd chip in with a simple, me too.
We have a bunch of HP ProLiant BL460c Gen8 and G7 Servers that reside in multiple HP c7000 Blade Enclosures and HP FlexFabric Virtual Connect Module . Veeam backup servers are generally HP ProLiant Dl380 Gen8. Backup server, 3Par V400 and 7400 are connected to fiberchannel switches. If we enable Storage Snapshots it works well in general but every now and then a few VM fails with unknows WWN. Almost never the same VM and no pattern to what LUN's eighter.

If we run it it without Storage Snapshots we get no errors at all so that's what we run in production.

[chjones]

So I did a test and we had another brand new BL460c Gen8 Blade Server that is destined for another purpose but I've managed to delay it for a few weeks so I could test it with Veeam.

If I configure a backup job to use that specific blade server, which can also contact the 3PAR via Direct-Attach through the FlexFabric Virtual Connects, the storage snapshots work with issue. Backups run fine, everything works.

I've had a few webex sessions with Veeam Support and they've been collecting logs and take screenshots and trying various things. No luck yet.

The good thing so far is that no one has tried to tell me "this is unsupported" when I explain how the ESX Hosts use Direct-Attach and FlexFabric vs the Proxies which connect via Brocade Fibre Switches, Everyone keeps saying "this should work", so that's a good thing.

[chjones]

If I configure a backup job to use that specific blade server, which can also contact the 3PAR via Direct-Attach through the FlexFabric Virtual Connects, the storage snapshots work with issue. Backups run fine, everything works.

Whoops .... meant to say "the storage snapshots work WITHOUT issue"

[emachabert]

Hi,

Using 7.0 patch 1, having the same issue, refering to a WWN of the 3Par array which is not shown anywhere in the management console.
I'm goind to open a support case.