HP StoreOnce CoFC SCSI Processor Device

[jchen522]

Is it normal to see close to 600 "HP StoreOnce CoFC SCSI Processor Device" under Other devices in device manager on a Veeam Proxy physical server?

When I look at how the zoning was done, there is one big zone with 3par, veeam proxy server, storeonce all zoned together.  Does it cause an issue? Does it slow down the backups?  The veeam backup gets to 99% but doesn't seem to finish its job.

[Mohammed_I]

Hi jchen522,

When using Catalyst copy over FC (only for COFC), the process is always bidirectional, so it is irrelevant whether you have an active-passive or active-active setup.

What you need is that all initiators are zoned in to all targets in both directions.

So assuming you have two ports Sa and Sb on the source and two ports on the Target Ta and Tb, you need to zone the initiators Sai and Sbi with the target ports on the target device Tat and Tbt and vice versa:

Sai with Tat, Sai with Tbt

Sbi with Tat,Sbi with Tbt

And to set up the reciprocal connection, the initiators on the target 

Tai with Sat, Tai with Sbt

Tbi with Sat, Tbi with Sbt

Assuming that port a and b are in different SAN, you will have less options.  In that situation you would end up with zonings for

Sai with Tat and Tai with Sat

Sbi with Tbt and Tbi with Sbt

If possible, always zone an initiator to multiple target ports so that in case something happens with one of the target ports, you still can set up a connection.

[jchen522]

Thank you for that detailed explanation.

When you said the following

"Assuming that port a and b are in different SAN, you will have less options. In that situation you would end up with zonings for
Sai with Tat and Tai with Sat
Sbi with Tbt and Tbi with Sbt"

Is this total of four zones or two zones? Single initiator, single target?

"If possible, always zone an initiator to multiple target ports so that in case something happens with one of the target ports, you still can set up a connection."

Is this single initiator, multiple target in one zone?

My veeam server has 1 HBA with 2 ports.  There are two SANs. There are two StoreOnce system each with 2 HBAs with 4 ports each.  In this case, I would create two zones or 8 zones ? which options below?

optional 1 - with 1 initiator to 4 target ports bidirectional in one zone for SAN A and another 1 initiator to 4 traget ports bidirectional in another zone for  SAN B?

optioanl 2 - with 1 initator to 1 target port bidirectional in one zone repeated 4 times to each 4 ports for SAN A and with 1 initator to 1 target port bidirectional in one zone repeated 4 times to each 4 ports for SAN B?

Thank you!

[yilmazaydin · Accepted solution]

@jchen522 Hello. Avoid combining all devices into one zone. Ideally, there should be one initiator in one zone. Also, it is not necessary to include storage systems of different levels in one zone. Below is a drawing showing the zones that describe your architecture. As one of the options.

Regards, YA.

[support_s]

Hello,

Let us know if you were able to resolve the issue.

If you have no further query, and you are satisfied with the answer then kindly mark the topic as Solved so that it is helpful for all community members.

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

[jchen522]

Does 3PAR have the capability to copy directly to StoreOnce? like RMC? If so, 3PAR and StoreOnce would be zoned together separately?

[CLDonohoe]

What version of Veeam Backup & Replication are you using?  There are reasons you could see the backup job "hang" at 99% for a very long time depending on the VBR version you're running, especially when using FC.

[jchen522]

Veeam is running version 11.0.1.1261

[yilmazaydin]

@jchen522 In Veeam environment t there is no way for devices to transmit data directly.  Therefore, there is no need to create a zone between them.

Regards, YA.

[CLDonohoe]

jchen,

You asked "Does 3PAR have the capability to copy directly to StoreOnce? like RMC? If so, 3PAR and StoreOnce would be zoned together separately?"  But you also mentioned you're using Veeam.

To clarify, Veeam has no ability to take any HPE Primary Storage snapshot and copy it directly to a StoreOnce.  For customers asking to do this, I routinely directed them to RMC when 3PAR and Primera were the prevalent primary storage arrays in customer environments.  Veeam always considered RMC a complementary product, because it fill niche use cases customers wanted.  It looks like RMC is still being maintained with it's latest release being in March 2024 - https://support.hpe.com/connect/s/softwaredetails?collectionId=MTX-MSC_RMC&tab=revisionHistory&softwareId=MTX_MSC_RMC36.

You also mentioned that Veeam is running version 11.0.1.1261 and that it gets to 99% and doesn't finish the job.  I strongly encourage you to upgrade to V12 to correct this issue.  In V12, Veeam introduced a new per-machine metadata format which increases the efficiency of job execution. 

TL;DR warning! 

The reason for the "hang" at 99% in V11 is that Veeam is performing an incremental backup for every VM in the job and then transforming all the backup chains at once to maintain a full restore point for every VM.  The reason it did this is there was once metadata file for the entire job.  So the job had to wait for every incremental backup to finish before it started the backup chain transformation on every backup chain in the job.  If you have lots of VMs in a job (which most customers do), then this would create a great deal of random I/O on the StoreOnce.  It was also misleading to the customer to report the job was 99% complete when there was a huge amount of work still to be done.

This is handled much better in V12 where individual backup chains have their own metadata file.  So as each incremental backup completes for an individual VM, the backup chain transformation process can begin immediately for that VM.  This can happen while other large VMs may still be completing their incremental backups.  There are also other performance advantages introduced in V12 like Catalyst fixed block chunking support, an improved Catalyst OS client, and support for Veeam gateway pools.

[jchen522]

I see.

@ CLDonohoe

Thank you for that valuable information!

[jchen522]

@Mohammed_I

Thank you for the detailed zoning information!

[Sunitha_Mod]

Hello @jchen522,

That's perfect! 

We are glad to know your concern has been addressed.