Typos in data loss prevention/recovery docs

nfriedly · January 18, 2017, 9:58pm

On http://rockstor.com/docs/data_loss.html#data-loss-prevention-and-recovery-in-raid5-6-pools

Suppose there is a raid5 pool called “mypool” with drives: sda, sdb, sdc, sdd. ssd is failed.

That “ssd” at the end should say “sdd”.

Also, In that same section, it says

A raid5 or raid6 pool in Rockstor requires at least 3 or 4 drives, respectively.

However, https://btrfs.wiki.kernel.org/index.php/Using_Btrfs_with_Multiple_Devices#Raid_5_and_Raid6 says:

Note that the minimum number of devices required for RAID5 is 2. In case of a 2 device RAID5 filesystem, one device has data and the other has parity data. Similarly, for RAID6, the minimum is 3 devices.

Is this a limitation of Rockstor, or another typo?

phillxnet · January 19, 2017, 8:26pm

@nfriedly. Hello gain; and yet more great finds.

You can see where the ssd typo crept in.

And the min disk spec is another doc error. However in another part of the docs: Redundancy profiles we have this right:

“Raid5: Two or more disks can be used with this profile, which supports striping + parity. Like Raid1, this profile can sustain a single disk failure. The BTRFS community consensus is that Raid5 support is not yet fully stable and so is not recommended for production use.
Raid6: Three or more disks can be used with this profile, which supports striping + dual-parity. Because of dual-parity, a Raid6 Pool can sustain up to two disk failures at the same time. The BTRFS community consensus is that Raid6 support is not yet fully stable and so is not recommended for production use.
”""

And the code itself also abides by these min drive specs:

github.com

rockstor/rockstor-core/blob/master/src/rockstor/storageadmin/views/pool.py#L347-L355


    handle_exception(Exception(e_msg), request)


if (pool.role == 'root'):
    e_msg = ('Edit operations are not allowed on this Pool(%d) '
             'as it contains the operating system.' % pid)
    handle_exception(Exception(e_msg), request)


if (command == 'remount'):
    return self._remount(request, pool)

And it’s funny you should quote that particular sentence of the btrfs wiki as early on, when that entry didn’t yet exist, I was unsure of this min drive spec and said as much in a pull request on the file I have quoted a section of above. As those figures seemed low. Anyway they were right and as a response it was actually our very own @suman who added that sentence to the btrfs wiki. Fancy that.

You can see our discussion over this on the associated pull request:

At the time this info was actually not that well known / available.

So hopefully the recovery section is the only place we have this wrong.

Do please open an issue detailing these required correction over at the rockstor-doc repo and if you fancy you are more than welcome to submit a pull request to enact them. You could also link back to this forum thread for context.

Thanks again for reporting your findings, it’s nice to have this issue revisited in this way. Also I was the last person to touch that doc file so I should have spotted it while I was there. Apologises for my not noticing when lasted editing that file.

Thanks for reporting your findings.

nfriedly · January 20, 2017, 8:34pm

PR submitted: https://github.com/rockstor/rockstor-doc/pull/162