I have read the guide on this and I have replaced data disks several times without issue. My situation is, I have 8 drive ports only (with all used) and want to replace the parity disk. In brief steps below? 

 

1. stop array
2. unassign parity disk

3. shutdown

4. remove old parity disk 

5. put in new disk

6. start up

7. preclear new disk 

8. assign to array as parity disk

9. rebuild parity from scratch? 

 

Question, if I have a data disk fail during the parity rebuild, would it be possible to put the old parity disk back in and rebuild the lost disk from the old parity disk? 

 

Have I missed anything in my steps?

Edited March 3, 20215 yr by SimonAG

Step 7 is redundant as building parity is going to overwrite every sector on the disk anyway.

 

2 minutes ago, SimonAG said:

Question, if I have a data disk fail during the parity rebuild, would it be possible to put the old parity disk back in and rebuild the lost disk from the old parity disk? 

Yes it can be done (but ask for the details if you need to do it) as long as no new data has been written to the array.

Well that saves loads of time. Will rebuilding the parity write anything to the array? I didn't think it would.

Just now, SimonAG said:

Well that saves loads of time. Will rebuilding the parity write anything to the array? I didn't think it would.

No - rebuilding parity only writes to the parity disk (and reads from the data drives).

OK thanks. 

I think these steps should be added to FAQ somehow. Today I’ve planned to write up the exact same question. Unfortunately there’s no clear guide for “simple” parity upgrade, only complicated parity swap instructions.

10 minutes ago, AnimusAstralis said:

I think these steps should be added to FAQ somehow. Today I’ve planned to write up the exact same question. Unfortunately there’s no clear guide for “simple” parity upgrade, only complicated parity swap instructions.

actually the steps in the online documentation for replacing a data disk or a parity disk after failure are identical.    Perhaps it just needs to be made clearer that they apply to both types of replacement.

11 hours ago, itimpi said:

actually the steps in the online documentation for replacing a data disk or a parity disk after failure are identical.    Perhaps it just needs to be made clearer that they apply to both types of replacement.

You're correct, but the lack of explicit mention of parity replacement confuses newbies like myself.🙃

Just to follow up, replacing the parity went through without a problem and took 25 hours to rebuild 40TB of data to parity whilst also going from 6TB to 12TB parity. 

 

I did however run in to a few strange problems. After rebuilding the parity, I was unable to delete anything from the array over SMB from windows, said I lacked permission. What could cause that? I am restarting again now. Just using the old parity 6TB to replace the final 4TB in the array and bringing total space up to 42TB.

 

I also had an issue with NFS after the parity rebuild, where it seemed to freeze up my vero 4k when accessing NFS shares. When I stopped the array vero 4k unfroze immediately, so it was like a stuck connection? It was working fine, i was watching content over NFS, then I noticed that some of the latest content was not showing up over NFS, from SMB in windows it was displaying fine. Then when I restarted Vero and went back to NFS shares that is when it started freezing up. I have since rebooted and will wait until this data rebuild has finished and then I will test the deleting over SMB and NFS from vero again. 

 

 

2 hours ago, SimonAG said:

Just to follow up, replacing the parity went through without a problem and took 25 hours to rebuild 40TB of data to parity whilst also going from 6TB to 12TB parity.

Just to clarify - this actually built 12TB of parity since that is the size of the parity disk.    It basically makes no difference how much data is being protected by this parity - the time to build parity is almost completely governed by the size of the parity drive.

Data drive replaced fine but still have the read only on windows over SMB, not deleted over nfs. Is there an issue? 

 

Any reason for me not being able to delete folders/files over SMB after a parity drive change? Do I need to reboot?

 

NFS seems fine now but not tested deleting from NFS.

Edited May 8, 20215 yr by SimonAG

From what I can see, it is only affecting files and folders that were shared to a filezilla docker, I have removed the docker completely but they are still read only. I can remove other files and folders from other shares. Strange one. 

NFS problems definitely caused by vero 4k and changing network config. So not related.

 

SMB delete problem i think is smb permissions as i can delete over nfs. it says tower/nobody lacks permission. i tried reexporting the smb. All shares are public so not sure the issue, Not urgent so i will wait for any advice before doing anything else. 

 

sorry for triple post.

Looks like I found the problem with the delete access. The Ownership of the files is set to user "nobody" instead of the user "windows username" which is the user for the rest of the files. Interesting that these files were created by the filezilla docker with the folder mounted in to the docker image.  Previously everything was downloaded through my windows PC to the NAS shares and that is why they end up with the owner "windows username" group "users" that allows me to delete those files. So a simple chown -R username:users might be sufficient here? I wonder if there is a way to force that user on to those docker downloaded files. 

 

So again completely unrelated to the parity change, just doing too many changes at the same time I guess

 

Chown worked fine so now just need to post in to the filezilla docker thread and try get those files to default to the correct user. otherwise have to setup a cron job to chown it i guess. 

Edited May 9, 20215 yr by SimonAG