• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Server room overhaul

M

Mobit

n00b
Joined
Mar 16, 2012
Messages
10
Hi folks,

Our current setup is less than ideal. I'd like to add some redundancy / failover to this system if possible.

At the moment, we have the following:

-1x Supermicro 4U server (SSG-6047R-E1R24L) housing 4 LSI 9286CV-8e controllers
-10x 24-bay JBOD with single SAS expanders (each with 1 input, and 1-2 outputs)

Each one of these JBODs is running in RAID60 (3x8) and we are hoping to continue taking advantage of the speeds we get running off these controllers as opposed to anything software based.

Archival isn't a concern fine as we move everything to LTO (twice, one off-site), but we want to make sure that should anything happen we don't have any downtime.

Is any of this gear still usable or am I looking at two new servers on top of new expansion chassis with multiple SAS expanders?

Thank you for any input you might have!
 
I would wonder what exactly your current setup is not providing.

You said you want to add redundancy or failover, but you have a bunch of internal redundancy within each of the JBODs, with the potential to survive up to 6 drive failures. Failover is something else and likely would be handled in software, mirroring the contents of some of the JBODs to another JBOD.

I would think you have a lot to work with already, but don't have a lot of info to go with.
 
I'm not worried about losing data, I'm more concerned with redundancy in terms of maintaining uptime. If we could figure out a way to keep everything up and running if in the event of hardware failure (These LSI cards run pretty hot... we've lost two in the last 5 years or so even at 68*F in our server room) or an OS crash that would be ideal.

Editing to add more info*

This area gets a tremendous amount of power outages, and despite our battery backups we don't always have someone staffed to quickly shut everything down. On occasion we've lost equipment due to this, and it's tough to come back from a weekend to find your staff is unable to work.

It's as TV production company.
 
Last edited:
Mobit said:
I'm not worried about losing data, I'm more concerned with redundancy in terms of maintaining uptime. If we could figure out a way to keep everything up and running if in the event of hardware failure (These LSI cards run pretty hot... we've lost two in the last 5 years or so even at 68*F in our server room) or an OS crash that would be ideal.

Editing to add more info*

This area gets a tremendous amount of power outages, and despite our battery backups we don't always have someone staffed to quickly shut everything down. On occasion we've lost equipment due to this, and it's tough to come back from a weekend to find your staff is unable to work.

It's as TV production company.
Click to expand...

Buy more batteries.
 
Firstly, you should be able to connect your UPSs to your computers and shut them down in the event that their batteries hit X percentage.

Secondly, if you have *that* many power outages, then at some point integrating a generator that will come online X minutes after power is lost (presumably before the batteries run out of juice) might be a consideration.
 
While I understand our staffing is not ideal, I don't think an additional two full racks of batteries is a viable option. Neither is a generator considering we lease a single floor of a 5 story building and don't exactly have the flexibility to implement something like that. The power outages have actually knocked out some of our UPS as well. Two of our systems won't allow any sort of automated shutdown without stopping certain services and backing up metadata, so I'm not sure how we would make that work. I could create a macro of some sort, but we do have additional reasons to opt for a high-availability solution. It appears there's nothing we have that will be usable going forward after looking into it some more, unless I missed something.

If you could ignore the why's for a moment, I'm open to feedback on the original question and would appreciate some additional input.

Thanks
 
I think true automatic-failover high availability is a big ask, and likely not something you can achieve with existing hardware; not to my knowledge at least.

You could ditch the server and JBODs but keep the drives and install them in something with high availability in mind from the get-go, like a pair of Synology devices plus a stack of expansion modules for each, or something equivalent. That would provide you with systems that could work in a HA cluster natively, and you could present the storage as iSCSI or NFS or SMB or whatever floats your boat. A 'real' SAN would be better, but no SAN will let you keep your own drives plus they are expensive as balls you might as well buy a redundant datacenter.

Beyond that, if you're OK with active/passive failover, you could do what you originally proposed - buy another server and set up some kind of realtime or near-realtime mirroring from your existing arrays to your backup arrays.

I think your most valuable asset in your racks is just the drives themselves, so your basic option sounds like 1) repurpose those or 2) replace those. If you decide to repurpose them, you'll need somewhere to put the data in the meantime (during the transition) and you might not have enough drives/space to do that, so buying new hardware may be inevitable.
 
You must log in or register to reply here.
Back
Top