Hi,
so I recently acquired both a ZeusRam and a small P3700 and decided to have some fun running a few benchmarks.
I have a pool consisting of 12 mirror'ed Toshiba DT01ABA300's which I have used for about 3,5 years now.
After running benchmarks on the pool for about a day I saw an email that one of the disks had a bunch of SATA errors.
Darn.
Ok, no problem, I have a replacement drive ready. So I plug it in, resilvering starts ... SATA error count on *newdev* increased ... Grrr.
Still no problem, I have a second replacement drive ready (I moved one of the drives to my desktop and didn't want to add another vdev and be without spare...)
Resilvering starts - SATA error count on *newnewdev* increased - now this is getting ugly :(
So I decided to move the third disk out of my desktop again where it was running fine (so I thought) and after much pain trying to clone it I just reinstalled the desktop to finally free up the third Toshi drive ...
I should have heeded the indications - as soon as I plug in the drive and resilvering starts... Pending sectors's on the third drive...
Needless to say that my offsite-backup had some issues recently (zfs snapshots not running as expected for 3 months and now its totally out of sync since I can't copy 4 TB over my internet connection in any reasonable time)
So, now I am looking what to do.
The pool was due for a makeover anyway, but that was not planned for *now* ;)
Options:
-Simply buy a new disk size 3TB+ and be done with it, better €80 less than loosing data (I'd rather spend the money on the new pool to be honest)
-Add additional protection to impacted vdev (only) to be reasonably sure I can keep the data save for a month or two until new pool is ready.
So after a long intro now the question - how would I go about option two?
I know I can have 2 or 3 or n-way mirror vdevs, but can I have 5 2-way mirrors and one 4way mirror?
This is my current pool :
What I would want is to add one or even two additional mirror drives to mirror-2 only because one of the drives in that vdev has shown a bunch of SATA errors. Those are not increasing under light read load, but I don't really trust it.
The other drives are fine, Smart is fine as well and no issues whatsoever (actually had next to no issues with these drives before).
Thanks:)
so I recently acquired both a ZeusRam and a small P3700 and decided to have some fun running a few benchmarks.
I have a pool consisting of 12 mirror'ed Toshiba DT01ABA300's which I have used for about 3,5 years now.
After running benchmarks on the pool for about a day I saw an email that one of the disks had a bunch of SATA errors.
Darn.
Ok, no problem, I have a replacement drive ready. So I plug it in, resilvering starts ... SATA error count on *newdev* increased ... Grrr.
Still no problem, I have a second replacement drive ready (I moved one of the drives to my desktop and didn't want to add another vdev and be without spare...)
Resilvering starts - SATA error count on *newnewdev* increased - now this is getting ugly :(
So I decided to move the third disk out of my desktop again where it was running fine (so I thought) and after much pain trying to clone it I just reinstalled the desktop to finally free up the third Toshi drive ...
I should have heeded the indications - as soon as I plug in the drive and resilvering starts... Pending sectors's on the third drive...
Needless to say that my offsite-backup had some issues recently (zfs snapshots not running as expected for 3 months and now its totally out of sync since I can't copy 4 TB over my internet connection in any reasonable time)
So, now I am looking what to do.
The pool was due for a makeover anyway, but that was not planned for *now* ;)
Options:
-Simply buy a new disk size 3TB+ and be done with it, better €80 less than loosing data (I'd rather spend the money on the new pool to be honest)
-Add additional protection to impacted vdev (only) to be reasonably sure I can keep the data save for a month or two until new pool is ready.
So after a long intro now the question - how would I go about option two?
I know I can have 2 or 3 or n-way mirror vdevs, but can I have 5 2-way mirrors and one 4way mirror?
This is my current pool :
Code:
zpool status main
pool: main
state: ONLINE
scan: resilvered 3.30T in 5h43m with 0 errors on Tue Feb 14 18:02:58 2017
config:
NAME STATE READ WRITE CKSUM
main ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
gptid/3370ec45-1a7c-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/36356fe1-1a7c-11e6-b6d4-000c2989f189 ONLINE 0 0 0
mirror-1 ONLINE 0 0 0
gptid/94740bc0-1a7c-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/9797151c-1a7c-11e6-b6d4-000c2989f189 ONLINE 0 0 0
mirror-2 ONLINE 0 0 0
gptid/b03ec4f1-1a7c-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/72897344-f2a7-11e6-94ed-005056844136 ONLINE 0 0 0
mirror-3 ONLINE 0 0 0
gptid/0b0df3cb-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/0db36f27-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
mirror-4 ONLINE 0 0 0
gptid/2325c1ad-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/254b5c2b-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
mirror-5 ONLINE 0 0 0
gptid/3bef9fe5-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
gptid/3e00141a-1a7d-11e6-b6d4-000c2989f189 ONLINE 0 0 0
logs
nvd0 ONLINE 0 0 0What I would want is to add one or even two additional mirror drives to mirror-2 only because one of the drives in that vdev has shown a bunch of SATA errors. Those are not increasing under light read load, but I don't really trust it.
The other drives are fine, Smart is fine as well and no issues whatsoever (actually had next to no issues with these drives before).
Thanks:)
Last edited:
