Jump to content
XPEnology Community
  • 0

Volumgroup with dissimilar HDD sizes


George

Question

Hi all

 

Currently have a volume group based on 4TB HDD, having some serious problems... off topic, to backup the data I'm about to buy some additional drives to backup to.

 

I require 8TB so easy is 2 more 4TB drives. other option is a 4TB and a 6TB.

 

Once I'm done with the backup/restore etc etc. I won't mind being able to add say the 6TB drive into the unit. how will this be handled though?

aka 4 x 4TB + 1 x 6B drive. = using SHR

 

G

Link to comment
Share on other sites

16 answers to this question

Recommended Posts

  • 0

The volume is already SHR... but then also the volume is current fck'd... had a single drive problem, tried to repair it, kept on initialising the replacement drive, then kicked if out and said drive crashed (brand new drive) eventually out of the blue the volume crashed, luckily it came back degraded after a reboot,

 

busy looking at what drives to order, to switch unit back on, and copy the data off. and then try and repair it.

 

at the moment looking at either a cheaper Baracuda, 8TB drive or 2 higher end Ironwolfs. but the logic atm says go with the cheaper and use to as a off line cold backup rather and rebuild the volume group 2 out what I think is 5 good drives. and remove that one drive that kept on causing a degrade.

 

G

Link to comment
Share on other sites

  • 0

for my education.

 

2 x 6TB drives will give me 6TB useable,

 

3 x 4TB drives will give me 8TB, as the 3rd is used as parity/block copies on a round robin fashion ?

4 x 4TB will give me 12TB ? approx

5 x 4TB will give me 16TB ? approx

 

thinking about the rebuild, to contain the blast radius if something happens again, will rather go with 2 volume groups, one for documents and photos, the amount of space required is approx 2 TB, so 2 x 4TB will be good enough, at the moment that leaves me with 3 IO slits into which I can fit 3 x 4TB drives, will need to get a pic card with additional ports, but looking at the doing a 3 - 5 drive volume.

 

G

Link to comment
Share on other sites

  • 0
7 hours ago, George said:

busy looking at what drives to order, to switch unit back on, and copy the data off. and then try and repair it.

maybe its not the drives, there might be other sources for your problems

when new drives fall out of a newly created raid it can be seen as a sign for deeper problems (its still possible that a new drive is faulty but it needs to checked in both directions, drive and system) - so what about logs?

/var/log/dmesg

/var/log/messages

 

4 hours ago, George said:

for my education.

 

SHR can be SHR1 = one drive redundancy and SHR2 = two drive redundancy

to pick up your 1st post with 4 x 4TB and 1x 6TB

it will be 5 x 4TB and the left over 2TB cant have a one disk redundancy as there is only one 2TB "piece", if you add 2nd 6TB disk a 4TB piece will be added to the 4TB array (a raid5 array) and the 2TB piece can form a raid1 (one disk redundancy) with the unused 2TB and the two raid's (6x4TB as raid5 and 2 x 2TB as raid1) will be glued together by lvm2 and you will see 5x4TB + 1x 2TB as usable date

over thumb its the sum of all disk minus the biggest disk of the system

 

Quote

2 x 6TB drives will give me 6TB useable,

yes

Quote

3 x 4TB drives will give me 8TB, as the 3rd is used as parity/block copies on a round robin fashion ?

not exactly, your wording might be the description of raid4? shr1 uses raid5 and the parity is distributed over all disks in that case

https://en.wikipedia.org/wiki/RAID_3#RAID_4

https://en.wikipedia.org/wiki/RAID_5#RAID_5

but resulting capacity is the same, number off (same size) disks minus 1

 

Quote

4 x 4TB will give me 12TB ? approx

5 x 4TB will give me 16TB ? approx

yes and yes but thats plain raid5 not SHR, if SHR(1)  is used with same size disks it will result in a raid5 and the content of that raid set will be put into a lvm2 volume and that is what you see as disk size, plain raid5 would be without that lvm2 volume

the trickery with SHR is you can later on add different type and size raid sets to the lvm2 volume (extending it), with a plain raid5 volume you would only be able to add a same size disk (and if the disk is bigger its simply unused and will stay like this for all bigger disks you add)

SHR is a clever use of whats already there, mdadm software raid and lvm2

but there are also things you can't do with SHR, like adding a smaller disk

to really do that you would need to destroy the whole set and start with the new smallest disk from scratch, that way the 1st raid5 created will be in the size of the smallest disk (even if its just one small disk)

like 1+2+2+4+4 TB disks will result in a raid5 of 5 x 1TB, raid5 of 4x1TB and a raid1 of 2x2TB, usable 9TB (remeber from above? sum up all disks and remove the (one) biggest)

Edited by IG-88
Link to comment
Share on other sites

  • 0

thanks thanks.

know Raid 5 well.

so in essence it's a SW based Raid 5 with this smallest disk (run per physical disk) usage. ending in creating multiple raid 5 groups, the groups then added together into a volume group.

 

As for is something else maybe wrong, been wondering that way also. thinking is first thing is possible the data ports...

It was really mis behaving last night when I was trying to copy the photos, volume would crash, I would restart unit, and it come back degraded, as soon as I started trying to copy again it would crash. leading curious me click on logs for driving crashing and it was showing plug in plug out plug in plug out... resulting m taking unit down from where it normally stand, opening up and reseating all cables. it's been running stable since then. firs priority for now is to copy all media off the unit. Will then go into how to fix it.

 

Adding some new data controllers is low impact, I only have 2 volumes, Volume 1 is a cache volume located on a M.2, Volume 2, this one is the problem, so if rebuilding it's easy to move this one... my docker builds, Plex cache/library it's all on V1, unaffected.

If this is deeper, like the MB... then worried, how do I replace the MB and keep my build/My NAS configuration. This is where I'm stick my hand up and hope someone can assist.

 

Still wondering about the better drive configuration to go for, V2 at the moment was build with 5 x 4TB drives, I had the 6th drive space, in waiting, allot help it was ;)

thinking of future rather going with 2 Raid5 groups, one as a mirror, ye thats really Raid 1 out of 2 x 4TB -> for documents/photos etc, as a new volume3 and then the remaining 3 (or 4 if I get an data controller) all reconfigured as Raid 5 into Volume 2 so that all the programs still find that media where it was expected. Moving the photos/documents is a low impact change that won't touch anyone, its just SMB shared out to one machine from where we copy onto the NAS.

the media is shared via the Plex and iTunes server so don't want to go reconfigured that, and have to re-index everything. (indexes are save on V1).

 

I'm just talking... still nervous, only had 4 hours sleep. working on this, and that included fighting with mosquito's and my son crawling into our bed mid night.

 

G

Link to comment
Share on other sites

  • 0

... Update...

 

So I've been able to copy all my data off the NAS. can either now try and fix the Volumegroup as it is/inplace. by removing reading the problem drive... or destroy and rebuild volume group into the 2 above Volume groups.

 

Although to do the above proper I'm going to move the one VG onto a separate pic based controller, which I need to order and have a 2-3 day lead time.

 

Thinking a 2 HDD group and a 4 HDD group,

 

comments.

G

 

Link to comment
Share on other sites

  • 0

Disk 4 is not being used and is ready to be added back into the array by selecting Manage in the Raid Group page.  Or, replace it and repair with the new drive.

The System Partition error is just telling you the copy of DSM on Disk 3 is out of date.  There should be a Repair option on the Overview screen and it will clear that up.

 

If you had a corrupted btrfs filesystem, I would advise you to burn that down and rebuild it with your backup.  If the only issue was the array becoming degraded, just repair it in place.

Link to comment
Share on other sites

  • 0

so if I click Raid Group, Manage, Repair, it then only allows me to pick the one drive currently not being used, when that drive (used drive 4) is not there, then the entire manage and repair is grey'd out.

 

On Overview, I can see repair, and this is what is then shown, with no further options.

 

The volume group was 5 drives, last time I hit repair it went crashed right after, It's as if I need to get the VG rebuild with the 4 good drives, remove the 1 drive, and once healthy... then extend it by adding the 5th drive back.

 

G

 

1209976401_ScreenShot2021-02-23at20_10_34.thumb.png.bb147f82d204db50aa717f8c3b22b959.png

Link to comment
Share on other sites

  • 0
16 minutes ago, George said:

so if I click Raid Group, Manage, Repair, it then only allows me to pick the one drive currently not being used, when that drive (used drive 4) is not there, then the entire manage and repair is grey'd out.

 

On Overview, I can see repair, and this is what is then shown, with no further options.

This is what is expected.

 

16 minutes ago, George said:

The volume group was 5 drives, last time I hit repair it went crashed right after, It's as if I need to get the VG rebuild with the 4 good drives, remove the 1 drive, and once healthy... then extend it by adding the 5th drive back.

You can't repair a RAID5 array with less than the full complement of drives.  You must have 5 drives to rebuild, period.

Link to comment
Share on other sites

  • 0

"You can't repair a RAID5 array with less than the full complement of drives.  You must have 5 drives to rebuild, period." this have to have 5 drives in Raid 5 would be a DSM impose minimum as it's not a Raid5 technology limitation. With this 5 drive Raid 5 limit, it sort of then puts a stop to my idea to rather re-create the disk into to VG's. one as a mirror pair and the 2nd as a 4 disk VG.

OR... are you saying as this VG was originally 5 drives, it now has to stay 5 drives, I can't have it reconfigure/rebalance itself into a 4 drive Rad 5 VG.

 

ok, going to pull this drive out, zero the original drive and see if i can get it to rebuild.

 

G

 

Link to comment
Share on other sites

  • 0
2 minutes ago, George said:

"You can't repair a RAID5 array with less than the full complement of drives.  You must have 5 drives to rebuild, period." this have to have 5 drives in Raid 5 would be a DSM impose minimum as it's not a Raid5 technology limitation. With this 5 drive Raid 5 limit, it sort of then puts a stop to my idea to rather re-create the disk into to VG's. one as a mirror pair and the 2nd as a 4 disk VG.

OR... are you saying as this VG was originally 5 drives, it now has to stay 5 drives, I can't have it reconfigure/rebalance itself into a 4 drive Rad 5 VG.

 

ok, going to pull this drive out, zero the original drive and see if i can get it to rebuild.

 

I'm saying you cannot use the UI to do what you are trying to do.

 

In theory it is possible to shrink volumes, array sizes and then restripe into a 4-disk RAID5 at the command line.  I don't recommend it.  Better to copy all the data off, delete all the arrays and rebuild it the way you want it.

Link to comment
Share on other sites

  • 0
On 2/21/2021 at 7:16 AM, George said:

If this is deeper, like the MB... then worried, how do I replace the MB and keep my build/My NAS configuration. This is where I'm stick my hand up and hope someone can assist.

thats not really a problem you could move usb and disks together to a different hardware and run it

as long as nic and storage driver are working its not much of a problem

918+/1.04b need at least a 4th gen intel cpu (hasswell or newer) but you can "migrate" to a different type like 3617 and run it even on older hardware, config and data will be migrated so no loss (just change the loader to 1.03b 3617 in that case and boot, synology assistant will suggest a migration, the change of serial and mac is no problem as most of dsm works without valid serial/mac)

if its ahci as storage and for nic a realtek 1G or intel 1G nic it sure will run

1.03b can be a little tricky to get running as it needs CSM mode when its uefi bios (and the non uefi usb boot device needs to be used) but in most cases thats a question of personal skill and experience, there are only very few newer systems not having csm/legacy mode and these usually run with 1.04b/918+ (and that can do uefi and csm)

 

Link to comment
Share on other sites

  • 0

... TMI ... 

 

if i can break it down, for myself, if it's a compatible MB, as long as drives goes back into same SATA slots numbers, and I use the same USB, all should just work ?

 

changing loader, well that starts sounding like a complete rebuild then, and "probably" loosing everything on my current setup which is something I really don't want.

 

but at the same time now might be the best time to get the loader and DSM upgraded, while wishing not to loose anything on my volume 1 DG.

G

 

😅

Link to comment
Share on other sites

  • 0

did some re-reading after 2 cups of coffee, things clearing up a bit, hazy ;)

 

with my older 1.02b loader on the usb. - see how you say move HDD's and USB together,

where are my apps installed, where is the configuration all saved.

 

and then the 2 VG's Volume 1 (which does contain data I don't want to loose) based on a M.2 and the 5 problematic drives in volume 2, which it looks like I will destroy and rebuild.

 

is there a how to do the loader / DSM upgrade, without loosing vg1.. and my apps installed.

 

G

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Answer this question...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...