Expanding a RAID1 array

[adrianjansen]

Hi All,
I have an AS-204 NAS running ADM 3.0.5.RDU1
2 WD Red 3 Tb disks as Disk 1 and Disk 2 in Volume 1 as a RAID 1 array
1 WD Red 6 Tb disk as Disk3, Volume 2, as a non-Raid disk ( irrelevant to this )

I read the NAS352 Online RAID level migration document
Got to the Volume Setup Wizard, "Replace the existing disks in "Volume 1" with larger ones.
The next step says "Remove the old disk and replace it..."
So I replaced Disk 2 with a 6 Tb disc, thinking that it was logical to replace the second one, and rebuild using the first one as source. Took a few hours to rebuild, and seemed ok, no errors.
Then got to the "Replace the next disk... and replaced Disk1 with a new 6 Tb disk. Again, wait a few hours, but after the sync was complete, got a message in Storage Manager that there was an issue with Disk 1.
This was all done with hot swaps.
Powered down, and reboot, and get a message that the system would restart from scratch, erase all data etc. WTF !

Replaced the two new 6 Tb drives with the original 3 Tb drives, and it all started up again properly... Phew !

So before I spend another couple of days on this, what precisely is the order of replacing the drives ? Does it matter ? , or have I done something else wrong ?

[Antonios]

Try starting up the NAS with ONLY the first used 6 TB HDD, the one that rebuild correctly, the one that was in slot 2.

Put it in slot 1, no other HDDs present.

No hot swapping etc, power down properly, remove all HDDs, then put that first 6 TB HDD in slot 1 and power up.

It should start as RAID 1 'degraded'.

Hopefully. Then, you can take it from there.

[adrianjansen]

Thanks for the suggestion. But it did not work. Replacing the Disk1 with the new 6 Tb drive and restart gives a message that the disk is not initialised, install a new system image, all data will be lost.
I take this to mean than the system ( ADM, I presume ) is installed only on Volume 1 Disk 1 ( drive bay 1 ) and the raid mirroring does not mirror the system partition, only the data partition.

So I guess I have to create a new system on the new 6 Tb drive, then follow the steps to expand the volume onto that drive. Pity it does not say that in the instructions about expanding the volume.

Is there then a way to clone the current system on the original Disk 1 on to a new disk, including all Apps and current network settings, to save having to set the whole thing up from scratch ?

[Antonios]

Right, it needs HDD1 or it does not start, not sure though if RAID 1 leaves anything out, this would prevent expansion and behave just as you described.

Either this or during the failed startup attempt with both 6 TB something was written to both HDDs (like a failed timestamp or something) that causes the startup procedure to fail with the supposedly healthy 6 TB on its own.

This issue causes some serious concern as if the former is true, then RAID 1 is a bit useless if HDD 1 fails, I sure hope they didn't f***k this up so completely.

I'm watching closely myself how this thing develops, as I am sure many others do, so please keep us informed as often as possible.

Now If I had this situation in my hands I would probably go for cloning the original HDD1 into the new 6 TB using another machine running some sector-to-sector cloning software and see how that would go.

If all went well and had my 6 TB both up and running, expansion and all, I would then run some serious testing using the old HDD 1 and HDD 2 to see if RAID 1 actually means anything or not.

Just my 2c.

[adrianjansen]

Thanks for that. I did some more, but ended up in even worse problems.
Re-installed the original 3 Tb drives, and booted up ok.
Took a settings backup ( from Backup and Restore ), just so I could set stuff up as it was now, without having to do a complete setup from scratch.
Power down and remove the Disk1 and Disk 3 Tb original drives.
Installed one of the new 6 Tb drives in slot 1
Powered up, and accepted the initialisation demand.
Wrote a new ADM image to the disk. Set up Volume1 as a simple 1 disk system.
Verified that it would boot - Yes.
Restored the settings from the backup file - ok.
System now looks correct with my groups etc ( but of course no data ).
Powered down
Inserted Disk 2 of the original 3 Tb disks in slot 2, thinking that I could copy the data from that back to the new Disk 1, just to get a start.
Disk 2 shows as inactive in the Disk Management. Cant find any way of making it active without initialising it.
Power down, remove both disks.
Re install the original 3 Tb Disk 1 ( which would boot before I did the setup with the 6 Tb disk ), into Slot 1
Power up. Will not boot ! - asks to initialise, WTF !!!
So now I have 2 disks, neither of which I can read the data from. This is worse than useless !
Even worse, I backed up some of the critical data ( but not all of it, because I ran out of space ) onto the 3rd 6 Tb disk in my original system, thinking that at worst I could get it back from there. This IS supposed to be possible, is it not ???
Now I boot from the new working 6 Tb drive in Slot 1, but the disk I backed up onto is nor readable either ( inactive ).
I give up, guess I just have to wait until Support gets back from holidays.
But I am not impressed, I thought a NAS was supposed to be pretty much fail-safe. This one is not !!

[Antonios]

Put these old HDDs 1 and 2 into another machine running Linux (could be from a live CD with some partition tools etc), you should be able to read them, see if partition tables are identical or not, mount partitions, etc. Both these old HDDs are supposedly healthy, so data is intact, only some flag or bad timestamp or whatever tells ADM that it needs to initialize them again, obviously this is not an option.

Can't figure why it wouldn't startup with the original HDD 1, try with both old HDDs 1 and HDD 2, if it starts with them both power down and up a couple times, then try again starting up with the old HDD 1 see if that corrects things.

Another thing if the above exercise works, try to rebuild RAID 1 with the old HDD 1 in slot 1 and the 'other' 6 TB HDD, the one that didn't rebuild right in slot 2, this would give you a full copy of data on that 'other' 6 TB.

TBH, best thing would be a 'break RAID 1' command, not sure if such thing even exists on Asustor, this would make HDD 2 independent of HDD 1 right from the start, both HDDs would have same data but would operate in 'single disk' mode.

Sure hope all ends well and also hope this whole mess is not another 'feature' of ADM > 3.

EDIT: Adding one last minute idea, could work for either of the old HDDs, use an external enclosure to see if NAS can read them when connected through USB.

[adrianjansen]

Thanks for your reply.
I dont have a machine running Linux. But I did in fact try installing and running Linux from a Flash drive. Not sure I have it right yet, although the desktop came up ok. And no, connecting either of the original disks to that machine does not work - the drives never show up, either as internal STA drives or externally over USB. I know nothing about Linux, so I dont know where to start there.
I did see something on one of the Asustor Knowledge base items that one should be able to do this:
https://support.asustor.com/index.php?/ ... using-a-pc
But I could not get the first command in step 9 "apt-get install madm" to work, could not find where to get it from. Have no idea what to do from there.

I will try again putting the drives in the NAS, and restart, but I already did it twice. Maybe its something to do with having by this stage created another single drive system on this NAS ? Is there flash in the NAS which maybe holds a record of what has been done, as a security measure ?

A "Break Raid 1" or even just a "Make disk active" command would be really useful !!!

I will try connecting one drive via USB to the NAS, but I am not hopeful.

I know ( now ) that I should have backed the data up externally first, but what then is the point of having a Raid system at all, if it wont survive even a simple disk replacement ?

I am sick and tired of this whiole mess, for something which should have been a straightforward procedure. If it wasnt that I have critical data on the drives, which I need to recover, I would throw the whole system in the bin.

[adrianjansen]

I did try connecting one of the discs to a SATA-USB bridge. NAS booted up with a single new disk. Connected one original drive to the USB. NAS Does not show anything connected.
Connected the same bridge and disk to Windows 10 machine. Windows Drive management shows the disk as having 3 partitions, but no drive letters. And of course cannot recognise the ext1 filesystem. This was more just a check that at least something can recognise the drive.

Filed a support ticket with Asustor. Will see what they say.

Meantime I will go and do something useful.

[orion]

I think the failed reason (your first trial) results from booting disk being disk-1 (newer 6-TB disk, but failed on the second RAID rebuild process). NAS should boot up from disk-1 which data is not correct. So you got the initialization wizard when reboot.

Here is the procedure that you can try:

• Boot NAS with your original disks (3TB * 2 with RAID-1, 6TB on disk-3).
• Make sure NAS volume is in sync-state. If not, wait for synchronization complete.
• Replace disk-1 (to new 6TB disk) first when NAS is still working. Wait for rebuilding process complete.
• After that, replace disk-2.
• After done successfully, go to ADM storage manager to expand volume-1 size.

Those process won't restart NAS. In case that there are something bad, NAS can still boot up correctly using disk-1 (if step-3 is successful).

[adrianjansen]

Thanks for that. But I got the situation where the original disk configuration would not boot. I put in a support ticket, and got excellent response fron a tech wizard, who managed to recover both the data off the 3 TB raid 1 volume, and eventually the 6 TB raid 0 volume. That will teach me not to rely too much on RAID
All good now, with backups on external USB drives