moving installation to slightly different hardware

[digity]

I'm running FreeNAS 11.0-U2 and I need to move to another motherboard which is the same make and model but older hardware revision. It's also moving into a new case with a backplane - the current case has no backplane, just SAS expander to hard drives (mini-SAS to SATA cables). Everything else is the same as they're coming over too. What's the best way to move this installation and keep everything intact (data, pools, shares, VMs, settings, etc.).

 

[m0nkey_]

Just move everything over, including your boot drive. When it boots up you may have to re-configure the NIC. That's about it.

 

[Arwen]

And of course, always make sure you have several off server copies of your configuration.

 

[digity]

And of course, always make sure you have several off server copies of your configuration.

I keep hearing this... does that simply mean the .db file generated when going to System -> General -> Save Config?

Sent from my SM-G955U using Tapatalk

 

[Stux]

I keep hearing this... does that simply mean the .db file generated when going to System -> General -> Save Config?

Sent from my SM-G955U using Tapatalk

Yes. Some set up scripts to email copies of the DB periodically.

 

[danb35]

Some set up scripts to email copies of the DB periodically.

...and some don't like sending the database, which contains unencrypted passwords, through untrusted mail servers.

 

[digity]

BAM! Came right up without a hitch (at least from what I lazily observed). I definitely didn't put the drives back in the same order, but the pool came up intact. I love non-Windows advantages!

In case folks/searchers are wondering about the details...:

Old build - Antec Twelve Hundred V3 ATX case - >Supermicro X8DTL-i hardware rev 2.01, 1 x XEON X5650, 8/16/24 GB memory(*), LSI 9211-8i -> HP SAS Expander -> 3 x mini-SAS to 4xSATA cables -> 10 x 2TB HDDs
New build - Nocro RPC-4224 4U rack mount case -> Supermicro X8DTL-i hardware rev 1.3, 2 x XEON X5650, 40/48 GB memory(**), LSI 9211-8i -> HP SAS Expander -> 6 x mini-SAS to mini-SAS cables -> Nocro RPC-4224 backplane -> 10 x 2TB HDDs


* = various CPU sockets pins got bent during previous troubleshooting and then a CPU upgrade - the damaged pins correspond to memory controller functions hence degrading total memory detection (detected memory kept changing during restart)
** = I bent some pins again during this current migration LOL

 

[Stux]

BAM! Came right up without a hitch (at least from what I lazily observed). I definitely didn't put the drives back in the same order, but the pool came up intact. I love non-Windows advantages!

In case folks/searchers are wondering about the details...:

Old build - Antec Twelve Hundred V3 ATX case - >Supermicro X8DTL-i hardware rev 2.01, 1 x XEON X5650, 8/16/24 GB memory(*), LSI 9211-8i -> HP SAS Expander -> 3 x mini-SAS to 4xSATA cables -> 10 x 2TB HDDs
New build - Nocro RPC-4224 4U rack mount case -> Supermicro X8DTL-i hardware rev 1.3, 2 x XEON X5650, 40/48 GB memory(**), LSI 9211-8i -> HP SAS Expander -> 6 x mini-SAS to mini-SAS cables -> Nocro RPC-4224 backplane -> 10 x 2TB HDDs


* = various CPU sockets pins got bent during previous troubleshooting and then a CPU upgrade - the damaged pins correspond to memory controller functions hence degrading total memory detection (detected memory kept changing during restart)
** = I bent some pins again during this current migration LOL

I've very carefully fixed bent pins in LGA1366 sockets before... very carefully :)

But damn... what are you doing to bend so many pins!

You don't push down or jiggle LGA CPUs when you place them in the socket...

 

[Evertb1]

** = I bent some pins again during this current migration LOL

https://www.intel.com/content/www/us/en/support/processors/000007101.html

You can do that 100 times without any bend pin.

 

[digity]

But damn... what are you doing to bend so many pins!

The hatches on those things keep giving me problems; I can't get leverage on the outside perimeter so I go inside to lift/open it up and my fingers ever so gently graze the socket pins.

Hmmm... I have LGA2011 CPUs coming in for another build... do those sockets have pins or pads? LOL... seriously, do they?

 

[Evertb1]

The hatches on those things keep giving me problems; I can't get leverage on the outside perimeter so I go inside to lift/open it up and my fingers ever so gently graze the socket pins.

Hmmm... I have LGA2011 CPUs coming in for another build... do those sockets have pins or pads? LOL... seriously, do they?

It has pads. Intel offers a wealth of information about it's sockets and CPU's. It really is not hard to find. And normally if you buy a CPU in the box it is also accompanied with a booklet. And Googling for photo's will help you out as well. And before you even come in the neighborhood of your processor, dry-practice with the socket so you know how it feels and looks to open and close the socket. I am building PC's since the time we had to deal with LIF and ZIF sockets (google them if you want) before Intel moved to the LGA style (officialy also a ZIF standard I think) and I never bend a single pin just by paying attention. And in the case of the LGA2011 socket: leave the protective cover in place while opening it.
http://www.pcstats.com/articleview.cfm?articleid=2732&page=3

 

[Stux]

I've only bent LGA1366 pins when installing upgraded processors (with heat spreaders) into systems which are designed for de-lidded processors.

 

[Ericloewe]

so I go inside to lift/open it up

Well, there's your problem. That's a recipe for disaster! You might even destroy the retention mechanism, which would be even more painful than a bent pin.

 

[digity]

People, people! It's okay, bent happens! LOL!

Sent from my SM-G955U using Tapatalk

 

[Evertb1]

I've only bent LGA1366 pins when installing upgraded processors (with heat spreaders) into systems which are designed for de-lidded processors.

There are actually systems designed for de-lidded processors? I always put that in the "crazy" overclockers category :). See what I know.

 

[Stux]

There are actually systems designed for de-lidded processors? I always put that in the "crazy" overclockers category :). See what I know.

The Mac Pro 2009 used OEM delidded processors. Well technically they were unlidded, not delidded ;)

This means a lidded processor was taller than the OEM version, so if you screwed the heat sink radiator down too far, you'd crush the socket.

2010 models were lidded.

 

[Evertb1]

The Mac Pro 2009 used OEM delidded processors. Well technically they were unlidded, not delidded ;)

This means a lidded processor was taller than the OEM version, so if you screwed the heat sink radiator down too far, you'd crush the socket.

2010 models were lidded.

OK, you live, you learn. I have never had the privilege to open up a Mac.

 

[digity]

OK, you live, you learn. I have never had the privilege to open up a Mac.

It can be quite beautiful actually. At least the towers and old pizza box models