Trouble Migrating from Single Disk to Raid 5

[adam3948]

I just got my Qnap 419P II and am having a bit of trouble with RAID migration. I looked around the forums but couldn't quite find someone in my circumstances so hopefully someone can help. I moved all my data onto a 2TB Single Disk on the NAS, then I had two additional 2TB drives that I put in the NAS thinking this hsould be just fine since the Raid 5 needs 3 disks. I went into disk management and and pressed the migrate button under single disk. Then I selected the two other disks that I had just added into the NAS. The migration started running and I expected to end up with all 3 disks in a Raid 5 Array with my data transferred onto them. The next morning the NAS was powered down and when I turned it on I had two disks in a Raid 5 Array and still one single disk with all of my data on it. What did I do wrong? Now I have a Raid 5 which I can't seem to delete. Should I have put in three additional disks to migrate everything to a raid 5 from the single disk?

[Postman99]

Bump ?

[P3R]

From your description it sounds as if you did it correctly. This is the documentation for Online RAID Level Migration.

A power down isn't natural (but could be caused by a power failure or having an unfortunate power down scheduled) when doing an Online RAID Level Migration so that happening is likely to have caused the migration failure.

If it still is a problem, contact Qnap support and see if they can help you save it. If not, I'd recommend a complete reinitialization into the disk configuration you want and then restore your backup data.

[SHoffEsq]

I have pretty much the exact same issue.

I have a TS-410 and had a mirroring RAID 1 with two 2TB drives (for a total available capacity of about 1.8TB). I wanted to add an additional two 2TB drives to the RAID 1 and migrate it to a RAID 5, but my RAID was listed as in degraded mode, so after unsuccessfully trying to fix it (along with problems logging into the NAS), I backed up all the files to an external drive, formatted my NAS drives, and reset the NAS.

After wiping the disks, I installed one drive into the NAS and set it up as single-disk. Then I copied all of my data back to the single-disk NAS. Once that finished, I installed the other three 2TB drives into the NAS and migrated the single-disk into a four-disk RAID 5 configuration (for a total available capacity of about 5.4TB). (I followed the migration tutorial at http://www.qnap.com/USEng/pro_features_RLM.asp exactly as described.)

However, my Web Admin page lists my RAID 5 (with the full 5.4TB capacity) in addition to the single-disk (with about 1.8TB capacity). Also, like the OP, all of my data is on the single-disk, not the RAID, so I have no data protection right now except for the external hard drive backup. Initializing the RAID 5 took about 40 hours, so I'm hoping I don't have to remove the RAID configuration and re-initialize it.

I'm going to contact QNAP Support about it, but if anyone else has any ideas or recommendations, I'd really appreciate it. Thanks.


Here are pics from my Web Admin:
Disk Management:
RAID Management:

[Postman99]

I have the same thing... a raid 5 disk volume drive (nothing !!)

Any ideas ?

[grocovid]

Many people seem to have problem migrating from single disk to Raid 5. And that's exactly what I would like to do in the next few days. I have a TS-412 with a single Seagate 3TB disk. I just purchased two additional 3TB disks, and hoped to get an extra 3TB of storage AND get a fault-tolerant system by migrating from single disk to 3-disk Raid 5.

According to http://docs.qnap.com/nas/en/index.html? ... gement.htm , I could go from "Single * 1 " to "Raid 5 *3" by adding two disks.

Yes but, the same page describes a procedure to go from Single * 1 to Raid 5 * 4 by adding not 3 but 4 disks , and I can't find a procedure for online raid level migration if I want to original single disk to still be used as part of the resulting raid 5 array.

Is that indeed possible? Isn't that the reason why the previous posters in this thread failed to do so?

Thanks for your help

PS:Same problem in viewtopic.php?f=25&t=54918

[P3R]

Yes but, the same page describes a procedure to go from Single * 1 to Raid 5 * 4 by adding not 3 but 4 disks...

The screen shots under Online RAID Level Migration unfortunately aren't consistant. But the exact number of disks used really shouldn't matter, the principle should be the same.

...and I can't find a procedure for online raid level migration if I want to original single disk to still be used as part of the resulting raid 5 array.

That's the main idea with all migrations.

Is that indeed possible?

Yes it is (or at least should be) possible.

Isn't that the reason why the previous posters in this thread failed to do so?

I don't know why the migrations described here have failed but it looks as if adam3948 had a power failure during migration. SHoffEsq initially had disk failure (that caused a degraded array) and if he wasn't successful in deleting all data from all the disks he had previously failed to migrate, that would explain his problems. These are all things that normally shouldn't happen and they may of course affect a migration in a negative way.

Make sure there are no active users on the NAS during the migration. Also disable all installed QPKGs, as much services and applications as possible. If this isn't done, these are other things that may cause problems during a migration or expansion.

A backup of data is always necessary and even more so when doing major system administration like a RAID migration. There is always a risk that something may go wrong, as we see examples on in this thread.

Personally I've done two successful migrations but that was from RAID 1 to RAID 5.

[prospero63]

To reiterate what P3R said, make sure you MANUALLY disable all services. What I found is if your NAS is a member of a domain you can't stop the MSFT file sharing functionality unless you remove it from the domain (local users only). It ** in that you have to manually recreate all the permissions, etc. but so far it's the only way I have been able to reliably do any migration operations.

[grocovid]

Can I leave read-only services on (like my NFS sharing, which is configured to be read only)?
So, basically, if I make full backups of my data, turn off al QPKG, turn off download manager and media server, turn off samba sharing, I should be allright? Do I have to turn off http(s) as well (for NAS administration) and only leave telnet/ssh on?

Anyhow, thanks for your help

[prospero63]

I turned off all of the "storage services". So NFS, Web Server (but not web management UI), MSFT F&P, iSCSI, download manager, media server, all QPKG, the whole nine yards. You might be able to leave read only stuff on, I'm not sure. Once I found a way to make it work I was more interested in getting it done so I didn't have to hassle with it again.

[P3R]

Can I leave read-only services on (like my NFS sharing, which is configured to be read only)?

I strongly recommend you don't do that!

It's not a matter of only reading or writing. All disks need to at some point be dismounted and then remounted during the migration process and something will at least not work as expected at that point. Worst case it messes your migration up and you'll join the others in this thread. The service interruption will definately be longer if that happens.

Now you're informed of the possible consequences, it's your call.

Do I have to turn off http(s) as well (for NAS administration) and only leave telnet/ssh on?

Web administration can be up but personally I like to play it safe and did only logon briefly to check the status. Maybe the reason my migrations have worked is I'm being overly cautious?

[grocovid]

Migration seems to proceed correctly. I am now at 51%. First step (mostly cp to the two other disks) took about 20.5 hours (going from a 98% full Seagate 3TB drive to 3x Seagate 3TB drive). Second step should take between 12 and 15 hours.

I would especially like to thank you for your word of almost paranoid caution. I disabled all QPKG, disabled all application servers, disabled all network services, finalized my backups (it's incredible how much non-backed up data you got even when you think that everything has been be backed up correctly !) and started the migration.

Then, since you had contaminated me with your paranoia, I checked that everything was indeed shut off using both top and netstat. I had looked to the top/ps report too quickly, but one thing that alarmed me was strange connections to hosts I did not know on the internet. In fact, my transmission qpkg was still running. Be careful: even if you disable the transmission qpkg (or the nzbget qpkg), your server might still be running. So, in order to shut things down properly, I re-enabled the web server, got into the transmission control web page, shut down transmission in a clean manner. I then checked netstat and ps : transmission was really gone this time. I tried to do the same with nzbget, with the server would not shut down: it was "sleeping". I just killed that process.

I am really afraid the migration might have failed if I had not shut down tranmission properly, since the original file system would probably not have been that easy to unmount.

I hope that this word of caution might help other migrating QNAP users (and that my migration will finish its course all right - I'm very confident about that one, it's only RAID-5 rebuilding at this point).

grocovid

[khyron_z]

Here is an article i noticed which if you read it says that in order to have the drives setup in raid 5 they dont all have to be the same drive type but when you set them up in a RAID -5 configuration they must have the same amount of empty blocks. So to answer this question you are having is you should be setting up the RAID-5 configuration with empty formatted drives. Which is probably why the original drive with data was not added in the RAID as it did not meet the requirements to be added to the RAID setup.

http://technet.microsoft.com/en-us/library/cc938485.aspx

Had a similar problem like this myself before and i found it was because i had data on the initial disk. Hope this helps.

[Remmerv]

I had the exact same thing as SHoffEsq above when migrating from a single disk to a 3 disk RAID 5 configuration of 3 WD 3Tb RED's: after the migration I ended up with still the original single disk in addition to a RAID5 of the correct size which was missing a disk (you will see that in the log file).
What helped for me was deleting the RAID5 volume (as the first disk is not included yet your data should remain on the single drive, it takes some time for the system to respond to clicking the remove button!), format the single drives 2 and 3 again so everything is ready for migration again. Then shutdown all services and packages to make sure no process is keeping the first single drive occupied.
Then perform the single disk to RAID5 migration again and on my 439proII it resulted in the required single RAID5 volume without any flashing red LED's afterwards! After activating all servers and packages again I could continue with the existing setup and data. All actions performed from the administration panel.
So it seems really necessary to stop all services and packages that could interfere with the migration process. As most people start searching the internet AFTER things are messed up (like me ) the above experience might be useful for getting things running correctly again, normally without loosing your data and from within the administration panel.
Still a good idea to backup before start though....