VSM build partitions

[chippa111]

Hi all

Currently doing a build for a customer that is co-located on a UCS box.

The server is going to have a total of 20tb for the media server to use.

I have been asked to split this 20tb into smaller 2tb virtual hard drives to be added to the server in ESX.

Is there any benifit to having 10x 2tb partitions on the VSM server that are all located within the same ESX datastore?

Or does it essentially provide the same result as having 1x 20tb partition assigned to the server?

Thanks guys

[Scott Olsen · Accepted solution]

The 2TB disk allocations are a technical limitation given the deployment environment.  I highly recommend you review the following documentation prior to deployment:

http://www.cisco.com/c/dam/en/us/td/docs/security/physical_security/video_surveillance/network/vsm/7_5/vm/VSM-7-vm-deploy-1.pdf

... Effectively, you are going to end up with multiple '/mediaXX' repositories on the VSMS instances, but that's okay.

Cheers,

Scott.

[chippa111]

Thanks Scott

I wasn't sure if it was a technical limitation or not.

Have set up the server with multiple partitions and all is working well.

Thanks for the feedback

Regards

David

[Scott Olsen]

Glad I could help!  We are a relatively small group here in the PhySec forums :-)

Cheers,

Scott.

[alexdrozdpro]

Hi Scott! We use VSOM 7.5.1 and encountered a problem with the video recording on the connected storage / media2. it appears in the system, but the recording on it is not happening. both sections / media1 and / media2 connected by iscsi-target VM hypervisor. what could be the problem? 

[Scott Olsen]

This depends?  How many video streams total are being recorded to this VSMS instance?  Was this media2 repo added *after* all the streams were in place?  

As far as I'm aware, recordings from an individual camera do not span multiple repositories, but whole cameras are balanced between available repositories.  (e.g. - If you were to add another camera to this system, it should be allocated to media2).  Is it possible this is whats happening?

[i.voevodin]

Hi Scott,

We use VSOM 6.3.1 and we also have simillar problem that alexdrozdpro describe.

On our server we have 2 /mediaX partion.

/media1 and /media2

We meet with crash one of our /media1, and we unmount this media and restore partition. After finishing all manipulation with /media1, we add this repositories to VSOM, but all cameras still writing archives to /media2.

From logs we view OVERSUBSCRIPTION groom action.

|2015-02-03 11:35:00| Volume /media1 OVERSUBSCRIBED at 98%, need to get to 95%
|2015-02-03 11:35:00|   OVERSUBSCRIPTION groom /media1/1001/1000/20150202
|2015-02-03 11:35:00|   OVERSUBSCRIPTION groom /media1/1001/20150203/54d0692a.smd
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1001/20150203/54d06a56.smd
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1001/20150203/54d06b82.tmp
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1002/1000/20150202
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1006/1003/20150202
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1007/1002/1000
|2015-02-03 11:35:01|   OVERSUBSCRIPTION groom /media1/1008/1002/1000

etc...

Could you please tell any possibilites to change on which mediaX cameras was writing archive?

[Scott Olsen]

:-(  I think I understand your issue.  I *think* Cisco has a script that can be used to "re-balance" existing archivers across multiple REPOs, but I don't believe I have this resource, and I can't recall where I had seen it before.

Hopefully someone from TAC or the PSBU sees this thread and can help out.

Aside from that, I think the only thing you could really do in the interim is stop some of the existing archives ('shelve' them), and restart a new archive for that same feed.  VSM should provision the new archiver on the media repo with the greatest available space (unless I'm mistaken).

[i.voevodin]

Hello Scott!

You exactly get my issue!

Thank you for recommendation, we try to request Cisco support for help with this issue.

I try your recommendation with stop archive and start again.

[Scott Olsen]

I think it's more than just stopping & restarting an archive that would be required to get your video spooling out to the new repository.

As a test, Just try creating an additional archive on an existing video feed.  The new archive *should* be started on the new repo. If it does, there's a way to proceed forward I can recommend, but hopefully Cisco TAC has an actual tool to accomplish it in a supported fashion that would be more elegant.

If Cisco doesn't have something useable for you to get 1/2 of your video transitioned over to the other repo, drop me a message here and I'll explain how I'd accomplish it if I had to myself.

Cheers!

[chippa111]

I had the same issue a while back when I added cameras to the VSOM when only 1 partition was partitioned.

I raised a Cisco TAC and they provided a script that can be run from the command line to balance the recordings over all of the partitions.

Hopefully something like this will just be added to new versions as standard.

[i.voevodin]

Hello chippa,

Maybe you could share this script to our community? )

Thank you!

[Scott Olsen]

I too, would like to see this script!  I think it was actually a posting of yours where I had first heard that such a thing existed. Thanks for speaking up :-)

Cheers!

[i.voevodin]

Okay Scott,

I got it! I'll try to test all suggestions in my lab environment and write here the result.

Best Regards.