NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
Maintenance needed before horizontal expansion of Flex-raid-1
I plan to upgrade my disks in NAD RND314 where I have two disk in FlexRAID-1 and one disk as JBOD. Run the latest firmeware. My idea is to just for safety right now incrase one 4TB to 8TB disk i...
janpeter1 wrote:
Sound good. Just a last question, I guess. Since I have one slot (out of 4) free I thought I could mount it there first and perhaps make some check, before I use it to replace one of the RAID-disks. Is that a reasonable idea? Perhaps make disk test of this new disk?
Personally I always test my disks in a Windows PC using vendor tools (Lifeguard for Western Digital; Seatools for Seagate). I run the long non-destructive test, and follow that up with a full erase / write zeros test. I have had some disks that pass one of those tests, but not the other - and I have sometimes found failures with just-purchased disks.
The NAS will do the short SMART self-test before it adds the disk to your volume. If you can't test the disk in a PC, then you could insert the disk the 4th slot, create a volume on it, and then run a disk test on that volume. Then destroy the volume, format the disk, and remove it. After that, hot-swap with the disk you want to replace. (Note that if you were running XRAID you couldn't do this).
FWIW, I suggest you reconsider your use of FlexRAID. You can make a full backup, and switch to XRAID (reconfiguring the NAS and restoring the data from backup). You'd have the same amount of storage as you have now - just on one volume. Expansion in the future would be a bit simpler.
Sandshark wrote:
StephenB wrote:
janpeter1 wrote:
I guess a "scrub" is a good idea to do on the RAID-1 pair before exhanging one of them?
I don't see the need (and wouldn't do it if I were in your shoes).
Ditto. A scrub from the GUI does a BTRFS scrub and an MDADM re-sync simultaneously. So, it's even more stressful on the drives than a re-sync for a drive replacement. If one drive is near failure, that might push it over the cliff and you'd know it's the one that you should replace first. But it could harm the volume when it did, or even push both over.
Plus the disk tests you just finished show that all sectors of the existing two drives can be read. Which is enough to sync the new disk when you insert it.
Hi,
Thank you very much for the input!
I understand SCRUB is a stressfull and long procedure longer that disk test.
Just to improve my understanding.
With two new and good disks, should I then see disc scrub in RAID-1 as that is the
time when possible bit-rot is eliminted? Ths bit-rot elimination is thus not there all the time right?
I have filled in a choice for the volume to eliminte "bit-rot" but the wording is different.
Does this switch only address the fucntionality of disc scrub?
Thanks
Bit-rot protection is a ReadyNAS feature that uses a combination of BTRFS checksums and RAID redundancy to repair files that silently change (e.g. experience "bit rot"). It's not directly related to scrub.
Scrub is a useful maintenance function - I do run it every 4 months on schedule, and I believe Sandshark also runs it periodically. I just don't think that you need to run it before doing your disk upgrade - the disk test you've already run is sufficient.
Sound good. Just a last question, I guess. Since I have one slot (out of 4) free I thought I could mount it there first and perhaps make some check, before I use it to replace one of the RAID-disks. Is that a reasonable idea? Perhaps make disk test of this new disk?
janpeter1 wrote:
Sound good. Just a last question, I guess. Since I have one slot (out of 4) free I thought I could mount it there first and perhaps make some check, before I use it to replace one of the RAID-disks. Is that a reasonable idea? Perhaps make disk test of this new disk?
Personally I always test my disks in a Windows PC using vendor tools (Lifeguard for Western Digital; Seatools for Seagate). I run the long non-destructive test, and follow that up with a full erase / write zeros test. I have had some disks that pass one of those tests, but not the other - and I have sometimes found failures with just-purchased disks.
The NAS will do the short SMART self-test before it adds the disk to your volume. If you can't test the disk in a PC, then you could insert the disk the 4th slot, create a volume on it, and then run a disk test on that volume. Then destroy the volume, format the disk, and remove it. After that, hot-swap with the disk you want to replace. (Note that if you were running XRAID you couldn't do this).
FWIW, I suggest you reconsider your use of FlexRAID. You can make a full backup, and switch to XRAID (reconfiguring the NAS and restoring the data from backup). You'd have the same amount of storage as you have now - just on one volume. Expansion in the future would be a bit simpler.
Thanks for your detailed advice here of testing the disk first and how to procede! I plan to do it on the NAS in the 4th empty slot, since I have no PC available to put the disk in.
When the testing is done and provided the disk ok, you think it is better to hot swap the disk to the RAID-1 volume I have? To me it sounds "safer" to turn the NAS off and then swap and let it boot with the new disk in the RAID configuration and then let it synchronize which may take a number of hour or a day or so. But still I understand that hot swapping is possible if necessary, perhaps even have some advantage over "cold swapping"?
About my configuration and reasons for it. The first volume two-disk RAID1 is for archive, photos, music, documents etc. The second one-disk JBOD volume is for backup of our computers in the home office. This volume in encrypted. We also do backkup of ther office computers on external USB as complement about twice a month or so. NAS take the regular backup and use macOS Time Machine here. So I a bit prefer to have separate volumes for this reason.
A (perhaps dated) technical reason to stay with RAID-1 is that it is what I understand technically easier than RAID with more disks and less redundancy. When I set the NAS up BTRFS was rather new for Netgear and I understood it as they took a bold technical lead here. Other NAS vendors have hesitated. Still at the time, here was uncertainty around how mature the "software" RAID in BTRFS was. Educated people told me that in professional use you really have hardware RAID and was sceptical here. So play it safe I stayed with RADI-1. Today, several years later I guess software RAID with BTRFS in ReadyNAS and perhaps also some other Linux vendors use it. I have not seen any alarming reports here, but I have not really looked for it either. Here is some recent talk from QNAP that question BTRFS snapshot etc but that is another story.
It is better to hot swap. That way, the NAS does not have to figure out what's going on when it boots with a degraded volume. You will see in the log that it saw the removal and insertion. NAS are built for that.
I have followd the detailed instruction by StephenB on page 13. So far everything goes as expected. I am now re-syncing and down 8 % in 6 hours. The original disk is 4 TB and it had about 3.3 TB data. This will take about 3 days. I just wonder if it is a good idea to just let it go on in a stretch - or good to shut down and restart next day, or so. Looking at performance heat looks ok 51 C for the new 8TB and 48C fo the old 4 TB and 41C for the almost in-active disk 3 that is also 4 TB.
Is it good to let the system rest a few hours - or is the sync then needed to start from scratch - more or less?
janpeter1 wrote:
Is it good to let the system rest a few hours?
Definitely not. Let it run to completion.
janpeter1 wrote: I have followd the detailed instruction by StephenB on page 13. So far everything goes as expected. I am now re-syncing and down 8 % in 6 hours. The original disk is 4 TB and it had about 3.3 TB data.Resync is resyncing the full 4 TB - mirroring every sector of the remaining 4 TB disk to the replacement. It doesn't matter how full the disk is - unused sectors are also mirrored.
janpeter1 wrote:
Looking at performance heat looks ok 51 C for the new 8TB and 48C fo the old 4 TB and 41C for the almost in-active disk 3 that is also 4 TB.
You can set the profile to "cool" - I sometimes do that when syncing (setting it back to balanced when done). Temps vary a lot in different disk models.
Thanks! Put it on cool and temperature dropped a few degrees. And yes you here the fan a bit. Before only in quiet mode, so a difference. I let it just go on now. Thanks again.
Mission completed! No problems. It all went faster when I increased the coolling and the whole sync of 4 TB to the new 8 TB disk took about 24 hours.
I understood that problems with a new disk turns up usually immediately or very early. I just wonder if my old former 4TB RAID-1 disk if there is any use to keep it for a while as is? I have sveral USB-backups so that is no real problem. Later I plan to erase this old 4TB disk and perhaps keep is a cold or hot spare.
The old RAID1 drive can be used to boot the NAS as it was immediately prior to it's removal if it was removed prior to any horizontal expansion. If it was after, it's not really going to help you with anything.
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
