It’s been a while since I have been to the forums, I like the new look. This system has been running since March of 2015(9.5 years not to shabby), and is currently running just to provide file services to the household.
System Specs
:
TrueNas 12.0 STABLE - U8
Supermicro X10SL7-F with XEON E3-1231v3
2*8GB Crucial ECC 1.35V DDR3 1600MHz
4 x WD30EFRX WD Red 3TB in RAIDZ2
2 x WD Blue 1TB Drives in Mirror
Boot Device - 2 x Toshiba SSD’s
EVGA 650w Modular Power Supply
I started getting Correctable ECC Errors being reported in the BIOS and messages:
The 2 WD Blue Drives are not serving anything at the moment the main RAIDZ2 Pool has 5 SMB shares on it but really only has one person using it at a time. Honestly the heaviest use of the system is probably during the CloudSync Tasks, or local backups to a USB Hard Drive.
I was curious on the experts opinion, if I was to remove the the 8GB module from the slot that is reporting the errors, would this system still perform enough to keep my wife happy, until I either build a new cheaper/TrueNAS system or migrate the data to a Synology/QNAP or some other solution.
If you have any suggestions or links to small builds that I could look at that I could consider building just to meet the requirements above that would be great as well. For the next 10 years I don’t foresee getting back into virtual machines and such, so a simple file server low power would be all I need.
Thanks to this community for my first build 9.5 years ago, it’s still running, except for this hiccup, and appreciate any assistance moving forward.
Possibly. But getting one or two DDR3 modules to replace the failing memory should not be too expensive. I do not see a need for a new system if you’re satisfied of this one.
I’d suggest, as a first step, pulling both sticks, cleaing the contacts, blowing any dust from the slots, then reinstalling. I think this MB has 4 memory slots, so you could also try the other 2 slots. ALso, I noted a new stick is ~$15 US so raplcing 1 or both is going to be far and away cheaper than any other option. Good luck. John
I agree with other: replacing the defective module (after have tested if is really defective as suggested), is cheap and fast!
And, offcourse, meanwhile you can remove it and use the nas with half ram (less ram for arc, but better then nothing )
FWIW, I’ve used 8GiB RAM for basic home SMB server without any issues. The users who get themselves in trouble with low RAM are the ones who also want to run VMs, apps (or in your case jails and plugins), or expect to saturate 10 gigabit.
So, no one is concerned about the 10 years on the motherboard? That is I guess my main concern. Everything else is replaceable, and thanks for price shopping the RAM, I didn’t even think to look, just figured as old as it was the supply was probably down and price would be high.
Appreciate the input, I will get everything backed up one more time and do some of the suggestions and see if it’s just a slot or a module or what.
Your motherboard will fail one day… but a failing motherboard would not result in ECC errors. RAM modules are the suspects.
Motherboards are quite replaceable: Just move your drives (and anything else that can be reused) to a new motherboard, boot, take a trip to the console to fix networking if necessary (i.e. if you used an onboard NIC), and you’re done.
Yeah, it was just a question do I say I got my 10 years out of this, time to get a new one in the house, and either run both side by side for now and prep for the failure.
I just don’t want to be the person that I constantly accuse others of being, you run your business on these machines and they haven’t been updated in 10 years…things fail. Just thinking ahead, or attempting to, but probably over thinking.
“ZFS is not a backup”, so getting a second NAS to serve as backup of this one is certainly a good idea—irrespective of the age of the primary NAS.
Updating to CORE 13.0U6 can be safely considered.
But in running a 10 year old system you’re in the good company of those here whose keep relying happily on a X10SDV motherboard… That would be the least of my concerns.
Okay, backed up everything, hauled the machine out and reseated everything, gave it a good cleaning, moved the DIMM to new slots and ran the upgrades that I had been slacking off on doing. No Mem errors reporting yet, see what happens as we move along here. I will probably run a MemTest overnight tonight and see what that generates.
Thought about starting another thread, but just thought that maybe I would ask here. I used to have the recycle bin working on these SMB shares and at some point it stopped working nothing was being moved to the .recycle folder by either user editing the shares. This is the configuration on the share:
Any risk to anything taking them out? I will be honest, I have no idea what these do or did or remember setting them, but they are on every share, some legacy thing carried over from 10 years ago?
No risk. The auxiliary parameter is actually actively harmful in that it breaks expected ordering of ixnas and streams_xattr (as well as breaking the recycle bin).
I’m not concerned about the age of my motherboard, I sometimes run stuff way older than 10 years. Whether to replace or not should be a decision driven by how much it would affect you and your home / business if your NAS is down for a few days while Wiredzone / eBay / whatever ships a new board to you.
For example, the (arguably!) best high-SATA count, low cost, low power, SOHO dedicated file server board SM ever made is still in production and you can buy it new or used. It costs about 530 at wiredzone and usually 1/2 at eBay. Search for X10SDV-2C-7TP4F. Techyparts on eBay recently was selling a truckload for like $250, which was low enough for me to consider buying one as a spare.
A better use of your time might be to consider what you’d want out of your next system (vs. present one) and then have a plan that you can move into action when the time comes for your old board to croak.
The whole reason eBay is replete with older systems being sold for dirt cheap is businesses making the decision you’re contemplating, ie casting off working systems because the warranty has expired, they’re getting older. But businesses measure downtime in $$ whereas most home users usually measure it in inconvenience. HTH.
)