VMWARE Backup with NBU 7.1.0.4

[Spartacus81]

Friends, 

Strange problem am trying to restore the vm to alternate datastore over the SAN and its failing with 2817.

after checking the zonning and everythign i reattempted the restore which sat there for quite sometime hence i cancelled it but in virtual center i can still see the netbackuap process creating vm? 

can someone shed some light?

[Mark_Solutions]

Is it a 2008 vmhost? Are its disks online?

Also is it thin provisioned - i seem to remember that also make a difference with restores

A few patches / hotfixes as well - are you up to date with everything for 7.1?

Going 7.5 / 7.6 soon?

Hope this gives some ideas

[Spartacus81]

sorry mark i cant seem to update the post as this community site is playing up.. 

Master 7.1.0.4 Linux

Media 7.1.0.4 W2K8 and also NBU vmware host

VMWARE backup method SAN

Backup over the SAN successful. 

Restore failing with 2817 and also killing the restore job doesnt kill the processes in virtual center..

yes vmhost is W2K8 and yes all vmfs disks/lun are online but not mounted (automount disabled)

restore is running but extremely slowly it looks like it hangs but if you see in vc you can see the create vm processes eventhough the nbu job cancelled 150. i have just logged a call with symantec. 

so basically what am trying to achieve here is 

a. is my restore running over the SAN?

b. backup took 40mins to complete with 60MBps throughput why restore taking that long as its going back to the same host but different lun?

c. why NBU doesnt cancel the processes in vc when you cancel the restore job in activity monitor?

thanks again and yes 7.5 is down the line. 

[Mark_Solutions]

I still get a 404 error on the Support site but if it is working for you then take a look through the 7.1 late breaking news to see if there are any VMware hotifxes that may help you.

Could we see the text of the detail tab of the restore job to get an idea of what it is doing please.

Other than that you would need the following logs for a VMware restore:

Master : bpbrm, bprd

Media /VMware host : bpbrm, tar, bpVMutil, bpVMreq

[Spartacus81]

had a look nothing relevant to the issue am facing.. 

job log is fairly simple - i will cut and paste when my media server comes online (intransit from DR site to prod)

am going to restore the data at production site using the tape library so does that tape libary needs to be connected to the FC or the vmware host (media server)? my assumption is as long as tape library is directly connected to vmware host via the FC it should give you the FC throughput? what are your thoughts mark?

[Mark_Solutions]

If the library is connected via FC and the VMWare host sees the datastore luns over FC then you should be OK - but dont zone them on the same ports - best not to mix disk and tape on the same FC port

So are you saying the host was on a different site before? Did it have FC connection to the datastores from there? If so how far was the link between sites and what sort of speeds to you have between the sites?

[Khajan]

Alternatively try restoring it via the NBD rather than SAN transport mode. I have seen it working. Good Luck.

-KG

[Spartacus81]

yes it did have the FC directly to the datastores - bit of a struggle with zonning but eventually got it working.

site to site link was 1GB and i was quite surprised to see the vmware backup throughput over the san..

[Spartacus81]

thanks khajan.. trying with NBD at the moment will share the result..

running a restore and creating a virtual machine procecss in the vc taking considerably long time to complete? is it transferring teh vmdk blocks at the same time or is it just creating a vm and will transfer teh data in a next process?