Stop Resilvering ?

[wtfuar]

Two of my WD Green drives messed up and started to go offline.
First 1 drive and I wanted to replace it but as I entered the pass phrase -with the massage this will invalidate the current key- it came back online and resilvering started -I did nothing, like swap the drive ... it just started to resilver.
After some minutes another drive went offline and I tried to replace this drive, but now my key is not valid anymore.
Finaly the already resilvering drive that went offline in first place, went offline again and the resilvering did not stop that I finaly could start swap this hazardous drives out.

I need to say that those 2 drives droped already before but after restart and mounting they always were back online and the volume could be decrypted and resilvered 5MB. ZFS is robust but I need to get those drives offline.

The volume is made by RaidZ2 (12HDD) & RaidZ1(6HDD) in Stripe
FreeNAS-9.3-STABLE-201509022158
3 x AXX6DRV3GEXP (unknown FW) 6x2GB Thoshiba, ABA 6x3GB Thoshiba ABA, 6x2GB WD Green EARX
M1015 IT-flashed P16

PS: Extended Offline S.M.A.R.T. tests show no errors on the disks anyway. So the plan is to throw all WD Green out and replace'em with Toshiba.
Thoshiba has error recovery and they run much more reliable. This is not a Server and does not run in 24/7 just a NAS - Network Attached Storage

 

[Bidule0hm]

The RAID-Z1 vdev is a disaster waiting to happen... You should copy the data elsewhere, delete this pool and make a new one with RAID-Z2 vdevs then copy the data back.

Plus it's not recommended to use encryption at the pool level unless you have a very very good reason to do so. Usually it's better to use a normal pool and use an encrypted container on this pool to store only the files that really need encryption (TrueCrypt, ...).

To answer the question, just let the resilver continue until it ends, it's the only way to recover a healthy pool.

Now I wonder why you have drives dropping one after another like this. Do you have setup SMART tests? Is your PSU decent quality?

 

[wtfuar]

I have the data backed up, thats safe.
PSU 300W was replaced with a 550W 3 weeks ago.
SMART extended offline ... shows all is fine.

 

[Jailer]

The remainder of your hardware would also be helpful in diagnosis. Also:

This is not a Server and does not run in 24/7

This is not a recommended practice with FreeNAS and ZFS. For ZFS to function efficiently you need to leave it up and running, especially with a large pool such as yours.

 

[FreeNAS_DIY]

May I ask a newbie question?

Why is it not a recommended practice to turn of FreeNAS?
I use mine only for backup purposes (although the NAS data is also backed up to an external drive) thus it is only a small amount of time turned on.

What is the main disadvantage of such practice?

 

[BigDave]

May I ask a newbie question?

Why is it not a recommended practice to turn of FreeNAS?
I use mine only for backup purposes (although the NAS data is also backed up to an external drive) thus it is only a small amount of time turned on.

What is the main disadvantage of such practice?

In order to monitor pool and drive health, the drives need regular smart testing and the pool should
be scrubbed to correct errors. If your machine is powered down, these very important procedures
connot take place. Also, it has been proven that hard disk drives (enterprise/NAS design) will last
longer if they spin 24/7

 

[Bidule0hm]

I'd add that each time you reboot you lose the ARC content so you have worse perfs than 24/7.

 

[wtfuar]

Thank you for all the additional information on running freenas/zfs.

To get back to my initial question: Stop Resilvering?

As to my research resilvering is a automated self healing process to restore faulty/destroyed data with healthy data available through redundancy if there is any.

This can’t be stopped manually as it is possible with a scrub –

Code:

zpool scrub –s poolname

In my case I had a problem with insufficient power supply resulting in R/W errors and hdd device drops that lead me to false assumptions.

All is back to normal now and I will write up my actual system through the time coming to help folks with similar problems.

I did not read any reasons on possible problems that could appear of not running 24/7 that are not already known and of course not freenas/freebsd/zfs related.

The impact on performance is more situational and more relevant on small data with very frequent access.
As I go with large amounts of continuous data streams I think there would be no noticeable impact. I’m correct here?