ONSSI Storage And Fail Over Questions

Can someone describe how ONSSI handles fail over? Details in terms of licensing etc.. Does the fail-over server need to have X quantities of camera licenses sitting there on idle?

For example, if I have a system that has 24 cameras all streaming back to a central virtual server – how does ONSSI handle a WAN outage with fail-over functionality? The fail-over server would be on the camera side of the WAN (obviously).

In terms of storage – can ONSSI use a NAS solution as a primary for 1500+ cameras? Or does it need block level read/write capabilities? Would ONSSI use a SAN as a primary and then archive off to a NAS?

Thanks.


I reached out to OnSSI to provide an answer.

Hi Mike,

I am a technical trainer with OnSSI.

I can answer some of your general questions, but for specific detailed design info, let's wait to see what our sales engineering team says (John reached out to them already) - they are far better equipped than I to provide answers to design-related questions regarding best practices for storage and archiving.

In your first paragraph, that's one i can knock out - licensing is per camera. The failover server 'becomes' the original (dead) recording server - as far as the system 'sees' it. i.e. no extra camera licensing is required because there are no 'additional' cameras. The cameras just think that the original recording server changed IPs and they send video to the new IP. (if surveillance cameras could think)

I'm not exactly sure what you are asking in your 2nd paragraph, but it is quite probable that your answer lies somewhere between pages 232 and 239 (Failover Chapter) in this document --> Ocularis ES/LS Manual. :)

For your last paragraph, you can record and/or archive to a NAS, but OnSSI does not recommend this practice (see 1st diagram from our training manual below).

For SAN (or other DAS) as primary with archive to NAS, there are apparently no disadvantages! (2nd diagram below) :) ...actually, I'm sure there are some disadvantages. Let's see what my team says... hold tight.

I also forwarded a message to our Tech Support Manager for input (his team sees the results of bad storage/archiving schemes), but he is out today. :(

Thanks for the responses. This is really helpful.

Thank you Marty for your excellent response.

Here are my thoughts on Mike's original question and Marty's response:

  • The Failover Recorder keeps a close watch on the Recorders it has been tasked to monitor, if a Recorder stops responding, the Failover will attempt to take over that Recorder's responsibilities, it will need to be able to create a network connection to the relevant cameras.
  • Writing live video directly to a SAN is perfectly acceptable (as long as the connection to the SAN is solid and reliable), and is successfully being done in many OnSSI Installations. The NAS should be used for archiving only.
  • If using the Ocularis ES/LS Recorders, it is recommended not to archive to the same drive as the one used for live recording.
    • There is no need to archive ever (archiving is most commonly used when video is being stored in location A, and then later on moved to location B (for long term storage purposes).
    • The Archiving process can be resource intensive (as a new database is being made out of the existing live database); it is for this reason archiving is not recommended if the video is not being moved to a different location.

Please let me know if you have any further questions.

Ari Robinson

Thanks again for the response.

As it pertains to NAS storage solutions, are there any enterprise VMS solutions that can use NAS as the primary storage? Virtual enterprise solutions consisting of 1500+ cameras.

I would think that all VMS solutions would need block level read/write capability.

Mike,

I think most VMS solutions can use NAS as primary storage.... but this method introduces a big fat failure point (BFFP) that requires 100% network up-time to avoid. The integrity of the database can too easily be compromised by any network outage. Most mitigate this risk by avoiding that BFFP altogether....

If anyone knows of specific VMS solutions that prohibit NAS as primary storage, please share.

Also, can some of IPVMs esteemed integrator members (or anyone else) weigh in with their thoughts on NAS as primary storage? Anyone do it? (and if so, why do you do it?)

If anyone knows of specific VMS solutions that prohibit NAS as primary storage, please share.

No less an authority than Alex K. recently stated Avigilon did not allow it, but the issue quickly became contentious, apparently.

Also, can some of IPVMs esteemed integrator members...

Third party esteemed only or is self-esteemed acceptable?

self-esteemed is perfectly fine.... :)

Also, as I am neither system designer nor integrator, I am curious why certain storage and archiving topologies are being used - and more importantly, avoided - out in the field.

I'm certain that somebody uses NAS as primary, and has some pretty good reasons why they do that... even if only for site-specific solutions.

Are there any NASophiles out there?

I had a customer who already was running a NAS for storage on a relatively small (75ish cameras) Genetec system. Honestly it worked "ok", but there were definitely times during heavy motion writing that any requests for recorded video would seem to be buffering too much. I think the customer originally put it in for the fact it was low cost and met a budget; but we eventually and recently converted them into a direct attached storage environment.