ECC vs non-ECC

[Stux]

I would look at your PSU and PSU wiring. Often the cause of random disk drop outs

Believe the 8TB disks put a strong load on a system.

 

[MR. T.]

The read speed can be slow due to the fact that related data may be anywhere on the disks. Meaning ZFS has
one concept of fragmentation, but SMR disks take that to a whole new dimension. This could potentially increase
read time dramatically, due to lots of disk seeking.

I see... i hadn't thought of that.

I have other disks on the NAS but since the main pool is made up of SMR disks i end up always reading from or writing to them.

I had a bit of time and was trying to find out why things don't work for me and now have 6 (of the 17) disks that are connected to the SAS expander working.
No idea why the other ones are not working... that's going to be my weekend... trying to find that out.

At this point i have a error message from when the disks drop off:

Code:

Device: /dev/ada1, unable to open device
The volume volume2 (ZFS) state is DEGRADED: One or more devices has been removed by the administrator. Sufficient replicas exist for the pool to continue functioning in a degraded state

Normally if i leave this as it is... more disks will keep on dropping off till the pool goes offline.
Double checking that everything is correctly connected (it always is) and rebooting gets the disks back online.
I never tried to reboot without going to check if i messed up somehow, but might try that tomorrow.

Any ideas on why this is happening? or how to get the disk back in the pool without a restart?

 

[Arwen]

I don't have a solutuion for your disk dropping issue, other than to try and use the following
ZFS command, zpool online agains the pool & drive(s). IF the on-line command works,
each disk would likely do a bit of re-silver to catch up to the current state of the pool. So you
would want to do one at a time, and wait til it's done before going on to another.

There may be a way to speed up your reads in the pool with the Seagate SMR Archive disks.
In theory, if the files are written one at a time, they would be less fragmented. Unfortunately
that would require a full backup, wipe the existing data, (write ZEROs to each disk), and then
re-create your pool.

Last, it's possible that the disks have something screwy set. Please supply the output of the
following command for one of the Seagate SMR Archive disks that have caused the problem;

smartctl -x

 

[styno]

Hi @MR. T. , quite an interesting/challenging setup you got there.

Not sure if you missed this question, but this is where I would start as well.

I would look at your PSU and PSU wiring. Often the cause of random disk drop outs

How do you power all this?