Raid 5 Degraded

[brk9422]

I recently had a RAID 5 (3 Disks) via ZFS working for several months without issue. After upgrading to the 11.2 branch and re-building the RAID from scratch I keep getting a degraded pool where only 2 of 3 disks are working (both Seagates). At first I had 2 WD Blue drives and 1 Seagate (all 3 TB) before I got the first degradation. After looking at the alerts and pool/disk status it said that one of the WD disks failed. I bought a brand new identical Seagate drive and resilvered the drive. There were no issues for a few days before it said that the other WD had failed. I ran a memtest and the WD HDD utility to check the "failed" WD drive, and no issues/errors were reported. When I run a df -ah from the shell they all seem to say the sector and disk size is identical, but am wondering if because of the different brands or w/e there might be an issue using different hardware. I will be happy to post any information necessary to troubleshoot this but do not understand what is needed to keep the RAID/pool healthy. Do I need to buy another Seagate (or WD, since the drives seem fine from the WD utlitity) or is there potentially something else going on? I have checked the SATA cable as well and can't seem to find any specific hardware failure but will be happy to try anyone's advice on this.

 

[Logikgear]

What is the smart data output for the drives?

Code:

smartctl -i /dev/sdc

where sdc is drive designation.
I have had drives pass the drive utility but not FreeNAS. The utility allowed "problems" if they are inside the fault tolerances. FreeNAS doesn't allow any errors without alerting you.

 

[Evertb1]

Please note that there are some rules/guidelines for posting on the Forum. Those rules are not here to annoy you but to help the other Forum members to answer your questions the best they know. So do yourself a big favor and read the Forum Guidelines. And then edit your post according to the guide lines. Another thing is that you should read the ZFS primer you can find it here. That way you wil be familiar with the correct ZFS terms and you don't need to use terms like RAID 5 anymore. It might help to get a more speedy reply on your question. Best of luck.

 

[brk9422]

I just noticed that this drive is 5400rpm while the Seagates are 7200rpm. Which would also explain why 1 Seagate showed errors when I had 2 WDDs running. Do you think that is why? If so, any way to get FreeNAS to support/handle this or should I just buy another drive so I have 3 identical ones at all times?

Thanks for the assistance thus far.

 

[Logikgear]

Having mixed matched drive speeds should cause errors that I'm away of. I have one 7200RPM drive in a pool with 5400RPM drives and haven't seen any problems in the last two months. I didn't have an extra 5400RPM drive laying around when I had a drive error. It looks like smart is available for this controller..

Could you run

Code:

 smartctl -a /dev/ada0 

 

[brk9422]

Sure, I attached it since the output is pretty lengthy. Hopefully it has what is needed since the disk is offline now. Any time I re-silver it goes offline automatically within a day if not less.

 

[Logikgear]

So that seems to look okay. Can you share a screen shot of the Alert that FreeNAS is giving you?

 

[brk9422]

I cleared it already...and did a bunch of disk swapping/testing so not sure if I can get the most recent. I will be glad to re-silver and force it back though if that would be best. I always end up with this as an end result:

 

[brk9422]

I just found an error for ada3 as well...and that is my boot drive. Not sure what is going on at this point...

 

[Logikgear]

What is your hardware Config?

 

[brk9422]

It is a PC turned server basically. The boot drive (1 320GB drive) is connected via PATA and then the ZFS drives are SATA.

 

[Night_Rider0]

I just found an error for ada3 as well...and that is my boot drive. Not sure what is going on at this point...

View attachment 28029

While the errors being flaged by the console are an issue to be rectified they pose no immediate threat to your array as this is just the boot device. I had a disk that liked to do this when i was using an old fashion raid card. when i switched to freenas i replaced the disk and it solved the issue.

 

[Logikgear]

What motherboard is in the system? model number?

 

[brk9422]

This is it here: https://www.biostar.com.tw/app/en/mb/introduction.php?S_ID=612 wouldn't be surprised if it is too old and time to upgrade.

 

[Logikgear]

So your getting error on your Boot drive probably due to it being PATA and VERY old. I would start with getting a small SSD for your boot drive like an Intel X25 32GB. Do a config back up then install FreeNAS fresh to the SSD then restore your config. Pick up another Seagate drive to match the other two and replace the WD blue. Pick up some new SATA cables while your at it. Using consumer drives isn't a great idea but a lot of people here do. I like to stick Enterprise or NAS drive personally.

 

[brk9422]

I don't mind doing that, and kind of already planned on it (minus the SSD, but prob well worth it as well). Even if just a guess (or a few), why do you think it is failing on the drives not being identical for the ZFS pool? Is it common to have issues with consumer drives that are not identical or could it be one of a million things that is going to be impossible to narrow down? More curious at this point than anything and to avoid similar issues in the future.

 

[Logikgear]

I would do the SSD mainly because the X25 32Gb &64Gb Intel SSDs are Enterprise and cheap on eBay ($25 last time I checked for the 32Gb). Their Reed/Write endurance is crasy high, around 1.2PB. So they last forever.

IMO, consumer drives aren't ment to stay power up and spinning like FreeNAS like to keep them for quick access. So they burn out (aka wear out) rather quickly. The Serial number for the one you did the smart data test on looks to be less the 4 years old. It should have lasted long. (Again IMO). Having drives miss matched in a pool isn't bad but probably not recommended. My guess on your last WD Blue is either a bad SATA cable or the controller board on the drive is going out. Worse case is the SATA controller on the motherboard is going out. But, not likely.

 

[brk9422]

Bought a new Seagate and the SSD. After installing the latest and greatest version of FreeNAS and booting up, I get this (below). Seems like the mobo is at fault after all??? I have another "server" laying around, but of course it doesnt have 4 SATA power ports so I will have to get an adapter for the old school 4 pin cables, or swap that out I guess. Let me know if this confirms my board should be taken to the shooting range.

It has been sitting at that screen (but keeps retrying) for the past hour.

 

[Logikgear]

Make sure the SATA controller is set to AHCI and not IDE in the BIOS. Also, when installing FreeNAS be sure to have all other drives detached.

 

[Logikgear]

Looking at the manual PDF, Once your in the BIOS.
It should like like this

Chipset > SouthBridge Configuration >

OnChip SATA Channel [Enabled]
OnChip SATA Type [AHCI]

Then try and install again.