one of the disk on the RN314 at my parents' home failed and I have just noticed that all disks there have >55k hours of on time, so... maybe it's time for some renewal 🙂
Oddly (but that must have been me) there are four relatively small disks (2x 500GB + 2x 1TB) in raid 5 configuration, but although I was able to make a second backup of everything also after the disk failure, I would feel more comfortable to put the two new disks in raid 1, since space is not really an issue.
If I remove all old disks and plug the new two disks I imagine that I will lose also the partition with the OS6 and it will be reinstalled starting from scratch -something that I would like to avoid.
So my question is, what should be the right way to do the work?
Should I remove the two 500GB disks (one of those is the broken one), plug the two new disks, create a new raid-1 volume on the new disks, then unplug the other old two? Is the NAS supposed to move the OS6 to the new ones for some reason?
I have made a backup of the configuration, but I would guess it's not going to contain also the configuration for the other applications running on the NAS, like Plex, VPN, etc., right?
Thanks for any suggestion!
Model: RN31400|ReadyNAS 300 Series 4- Bay (Diskless)
If I remove all old disks and plug the new two disks I imagine that I will lose also the partition with the OS6 and it will be reinstalled starting from scratch -something that I would like to avoid.
Normally the process would be to replace the disks one at a time, and wait for the system to resync before moving on to the next. But that would require you to replace all four.
The other way is to back up the files, power down the NAS and remove all four disks (labeling by slot). Then put in the two new disks, and power on the NAS, doing a factory install. Reconfigure the NAS (creating any user accounts, shares, installing any apps, etc), and restore the files from the backup. OS-6 will be reinstalled from scratch (as you note in your post). Is there a reason why this is a problem?
There's no way to shift from 4-disk XRAID to RAID-1 via the web ui. There is a fairly risky way to do that with ssh/linux command line.
Personally I tend to run the disks until I see some signs of failing. Though of course others will replace them on a schedule.
Well, one died from one moment to the other, with no previous advice. Now I can't even see it's details any more, so it's dead dead. I see that they all have the same age, so... I prefer to avoid disasters. Even if I have backup (I have learned my lessons).
Normally the process would be to replace the disks one at a time, and wait for the system to resync before moving on to the next. But that would require you to replace all four.
The other way is to back up the files, power down the NAS and remove all four disks (labeling by slot). Then put in the two new disks, and power on the NAS, doing a factory install. Reconfigure the NAS (creating any user accounts, shares, installing any apps, etc), and restore the files from the backup. OS-6 will be reinstalled from scratch (as you note in your post). Is there a reason why this is a problem?
Well, one is that I have to do all the process remotely connected to one of the PC, the other is just... time. It's not my NAS, so I need to find out how I configured everything long time ago and redo it.
There's no way to shift from 4-disk XRAID to RAID-1 via the web ui. There is a fairly risky way to do that with ssh/linux command line.
Ok I think I might have not explained myself properly here. It's not that I want to shift from XRAID to RAID-1, it's more like, I would like to get rid of XRAID and start a brand new RAID-1 without being forced to do a factory reset because taking all the disks off would mean to lose the copy of OS6.
I don't know if it's a bit more clear in this way...
You can save the configuration from the System/Settings page. Then restoring it on the newly factory defaulted NAS will take care of everything except the actual data and apps. If there are any apps, you do need to restore them before restoring the config backup. In the long run, that will be cleaner and faster, since you already have the data backed up for restoration. Oh, and if you use user home folders, the user must log in via ReadyCloud or SMB (depending on the type of user) for the user folder to be created. Then you can restore that data.
You could leave the RAID5 degraded to leave an empty bay to be filled by a new drive on which to create a new JBOD volume, then destroy or export the original volume, remove the remaining old drives, add the second new one, then add redundancy to the JBOD to get RAID1. But that process has many potential pitfalls which could push you into the first scenario to recover. As with starting over, destroying or exporting the main volume does not move apps though it is supposed to (and usually does) move their location to the new volume. This option's only plus is that you can copy data from the original RAID directly to the new one instead of using an external backup.
For the future, you may want to look into a VPN (I use ZeroTier, which is not a traditional VPN) to make it easier to do remote administration. You could then log directly into the NAS instead of going through a PC on the remote network.
Thanks for the suggestions, I'll see what I can do.
You know, my biggest issue is that it's not the NAS which I use every day, many things were configured years ago and getting all stuff back -particularly apps- will be a very tedious job.
Just for my understanding, let's say that I don't care about the data (it's all backed up twice), if I remove two disks and leave the RAID5 degraded, then plug the two new drives, then create a new volume on those two new disks, then remove the last two RAID5, do I expect that the OS6 copy is transferred on the new disks?
if I remove two disks and leave the RAID5 degraded, then plug the two new drives, then create a new volume on those two new disks, then remove the last two RAID5, do I expect that the OS6 copy is transferred on the new disks?
If you remove two disks the volume will be dead, not degraded. Since the data volume holds your apps, you will then need to reinstall them.
What you can do is
uninstall your apps
switch to flexraid
destroy the volume
remove two of the disks
replace the disks with the new ones.
create a new RAID 1 volume on the new ones
re-apply the volume settings you want (quota, maintenance schedule).
reinstall the apps
recreate your shares and any backup jobs.
restore your data from the backup.
The OS is preserved. But at the end of the day, you won't save any significant time or effort compared with simply starting over. You avoid needing to set up the user accounts again, and perhaps setting up the NAS network configuration settings. Those setting are easy to re-apply for almost home NAS users, since there usually aren't very many user accounts.
The simple way to accomplish what you want is to replace all four disks one at a time. That eliminates any need to change the NAS setup. The extra work here is because you want to start off with RAID-1 in the new volume.
Either way, if something were to go wrong, you need to start over anyway. So you will need to document the configuration before you start.
As @StephenB said, you can't remove two drives, and that's not what I said you could do. You can remove the one bad one, leaving the RAID 5 degraded (but not dead, as it would be if you removed two) and gaining one empty bay. The Apps are on that volume, not in the OS partition, so you need it. You then switch to FlexRAID, insert one new drive, and create a JBOD volume on it. After copying whatever you need from the orignal volume (including apps, using the process in the linked post), you can destroy or export the old volume. This will put a copy of the OS partition, and all your settings, on the new volume; so you don't need to restore them. Then, you remove the remianing old drives, add the second new one, and use it to add redundancy to the new JBOD one, converting it to RAID1. The destruction or export of the primary volume should cause the NAS to designate the new volume as the location for apps and home folders, but will not move any content from them. I say should because I did run a set of experiments and had that fail to work once. I do not know why, and I was running a lot of experiments with that NAS along with that one, so it may have been an unusual combination of actions that caused it. I often use drives with some errors as my "sandbox" drives, so that also could have been the catalyst.
If you EXPORT instead of DESTROY the original volume, you will be able to put it back in the NAS together with one of drives of the new volume if you find you forgot something. Of course, that new volume will become degraded if the RAID1 conversion has already happened, and will have to re-sync with the other drive once it is re-inserted. And you need to EXPORT the original again before removal.
If you simpy remove the old volume (with power off) without exporting or destroying, then you destroy it when it is not actually present, you will be able to boot the NAS with just those drives, but cannot add them while the other volume is present. BUT, I do not know if that will properly transfer the home folder and apps location to the new volume, since I have not done it with a primary volume.
During this process, your RAID5 is non-redundant, so failure of a second drive kills it. So make sure you have done all you can (backed up data and configuration and recorded any app settings you many need to restore) in case you end up having to start from scratch with the new drives anyway. But one thing this does not do is make you have to re-sync your volume with it's old drives, which is definately riskier (how much depends on how close to falure the drives are). The OS ands swap partitions will RAID sync, but they are small.
Once you have just the new RAID1, you should be able to switch back to XRAID. But if you do need to re-insert and import the orignal (exported) volume with it, you'll need to go back to FlexRAID.
Thanks Sandshark, I'll give a look at the post you are suggesting.
So, do I understand correctly that the applications installed in OS6 are not stored on the OS6 partition? I thought they were staying in there, so if you manage to save the OS6 partition, you also keep all the applications.
The /apps folder in the OS partition is simply a symbolic link to a folder on the main partition so as to keep (properly behaving) apps from filling the smallish OS partition. You'll see how it all works as you work through the transfer process I posted.
There can be a few app configuration items (like modifications to .conf files) that have to be in the OS partition,. which is why apps need to be restored before restoring a backed up configuration. If you follow my transfer process, that won't concern you, but I'm stating it for anyone else reading this post for guidance.
I see...that is why I was assuming that they were on the same partition, I did see some configuration file there in the past, so I thought they were installed there.
Ok, thanks again!
I'm starting to get convinced that a clean start is better 🙂
I see...that is why I was assuming that they were on the same partition, I did see some configuration file there in the past, so I thought they were installed there.
There are a lot of mount points onto the OS partition. If you want to look see it w/o them, you can enter
# mount --bind / /mnt
Then look at .mnt.
Undo this be first navigating out of /mnt and then umount.
