Backup exec 2012 and Path Failover

[Willie_SA]

I need to know if BE2012 supports path failover? That is if I configure redundant FC links to the tape library. I will install EMC PowerPath on the host OS (MS 2008 R2 64bit) which would manage the HBAs on the server and manage the active/passive failover switching.

[Kiran_Bandi]

I am really doubtful about PowerPath managing redundent paths to tape drives. It do so for CLARiiON and Symmetrix LUNs. 

And if you don't mind, can you tell us about the connectivity a bit. How many drives in tape library and how the tape drives are connected to backup server?

[CraigV]

Hi,

PowerPath doesn't work with tape. You'd zone the server for multiple tape drives etc. Those paths aren't MPIO-capable, so if a path goes down, the drive goes offline within BE.

No way around this unfortunately.

Thanks!

[Willie_SA]

Hi,

I gathered more information regarding the configuration. Here is the comlete configuration - so the question remains - how will BE handle the failover between the library ports. Will the services/server need to be interupted, scanned and maybe booted to accomodate the switch of the channel

Thank you

VMware on 2 x HP servers attached to fibre SAN switch

EMC VNX with SAS and Nearline SAS (for d2d) also into the fibre SAN switch

Standalone HP server with MS 2008 R2 64bit for Backup Management, also HBA linked into the SAN switch. Will have dual redundant HBAs

Now, I will Zone SAS disk space to the Standlone HP server for disk storage (it will also run the anti-virus app), and also zone NLSAS for D2D backups, and also zone the tape library (which is also SAN switch attached, with 2 fibre cables). It has a single LTO5 fibre channel drive, but 2 FC ports in active/passive config. So path will be:

Server -> HBA1 -> switch 1 port 1 -> library FC port A

Server -> HBA2 -> switch 2 port 1 -> library FC port B

EMC Powerpath installed on the server. It manages the active path, so that the HBAs are in active/passive mode. You can’t have them both active as it will cause SCSI ID issues for the OS and you will see the same LUN twice in Windows Explorer!

Anyway, so if switch 2 port 1 goes down, the path between server and library is broken, but there is a redundant path and the EMC powerpath software will make HBA2 the “active” card and thus bring up the new path and operations can continue.

What I needed to know was this: does BE support path failover? If this happens during a backup operation, will the operation continue along the new path, or fail? Or even while the system is not in a backup job, will it lose connection to a SCSI target device if the path changes? Will one need to reboot the application to pick up the new path again? In Windows, that is not required. The device remains available. Don’t need to scan for devices again.

[Kiran_Bandi · Accepted solution]

Quoting from HP Document: "When properly configured the change is minimally disruptive to thehost and does not require any configuration changes on the host or in the backup application.If no drive commands are outstanding when a failure is detected, then the port change happenswith virtually no disruption to the SAN. If a command is outstanding on the link when a failure is detected the drive is not able to recover the command so that command will fail but the applicationwill be able to continue to use the drive on the new path. Many applications are able to recoverfrom a single command failure as long as the communication path to the drive is not lost."

See pg no 28 in this document: http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c02871554/c02871554.pdf

BE supports data path fail over. 

Broken Link: Drive 2 continued to operate without pause. Drive 1 activity paused for about 5 minutes following the link break and then resumed on the second port. The backup on both drives completed successfully and the application was able to return the volumes to the source storage locations. 

[CraigV]

@Kiran: I've had an HP MSL6060 lose connectivity to 1 SAN switch and BE never failed over, even with the server configured as such. The drive went down in Windows, and consequently in BE as well.

[Kiran_Bandi]

OP is confirming "In Windows the device remains available. Don’t need to scan for devices again."

So, i hope it will work. And HP document i quoted in my earlier post describing most/all of the cases.