Snapshots not showing consistently in GUI RN102 FW 6.10.3

[flatpack27]

Hi

I havean Rn102 with two 2TB disks running as a RAID1 mirror.

I've been doing some tidying up and deleted a large chunk of files but the space hasn't been released. Files were deleted about 18 hours ago.

If I look at the GUI it shows a block of the disk spaced used by snapshots but when you look at the listings on shares on the GUI no snapshots are found. Equally when you select the snapshot option on shares nothing is shown.

I have seen other posts suggesting using a couple of commands from the shell using SSH but was concerned as I don't have much free space at the moment so is that the way to go?

Would the following commands likely resolve this:

# touch /.force_snapshots_upgrade
# systemctl restart readynasd

but with the lack of free space could I cause myself a problem?

This is a screenshots of the GUI, do have another shot of the snapshot page but doesn't look like I can attach more than one file.

Thanks for any help and advice

Cheers

[StephenB]

One thing is that you let your volume get much too full.  You can run into big problems if you don't have at least 10% free space with btrfs, and really ought to keep free space at 15% or more. An OS-reinstall won't solve your problem, and I don't recommend trying that (either with ssh or the boot menu).

My first suggestion is to back up the data (if you haven't done that already).  Your data is at risk, and there's a pretty good chance that you'll need to rebuild the NAS from scratch and restore data from backup. 

Next, log in using ssh. Make sure you log in as root (NOT admin).

Now see if there's any possibility of getting to a reasonable volume fullness.

# btrfs filesystem df /data
Data, single: total=75.81GiB, used=64.44GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=15.87GiB, used=8.84GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

The part in color is the response of course, and yours will have different info than the sample above.

The part in red is the important bit.  If the amount used is more than 1.6 TiB, then you need to either delete more files, or upgrade to larger disks, as you can't get to 10% free space with your current disk(s).  Backing up your data and starting over with larger drives is the best option. 

Assuming you have < 1.6 TiB used, you could try running a balance from ssh.  Make sure you log in as root (NOT admin).

# btrfs balance start -dusage=0 /data

That will reclaim any completely empty btrfs chunks.  The btrfs chunk size is 1 GB - and you have only 180 MB of free space reported.  Hopefully there are some completely empty chunks to be reclaimed. Note this runs in the background, so you will need to wait for it to complete.  The status command below will tell you the progress.  

# btrfs balance status /data

If that frees up enough space to get you back to at least 10% free space, then try running a balance from the maintenance options in the web ui.

If it frees up some space, but not 10%, then run the balance start command again with a higher -dusage threshold.  I suggest starting at dusage=10, and then increase it in steps of 10 until you get to over 10% free.  After that run the balance from the web ui.

If you can't free up the space using this approach, then you will need to start over with a new volume.

[flatpack27]

Thanks for the information

The figures from my unit show for data total and used to be the same unfortunately.

I'd deleted 448Gb of data yesterday as I was doing a clean up but that hasn't appeared on the system as free space.

If I go into the .snapshot directory for the share I can 52 directories named c_10digitnumber. Some go back as far as December 2013? So I guess these snapshots are stuck so is there another command to allow them to delete?

The system logs show some snapshot delete activity earlier in April by snapshot prune worker and earlier in March some snapshots were deleted due to low free space.

Thanks for your help

[StephenB]

---

@flatpack27 wrote:

If I go into the .snapshot directory for the share I can 52 directories named c_10digitnumber. Some go back as far as December 2013? So I guess these snapshots are stuck so is there another command to allow them to delete?

---

You could delete the snapshots using the btrfs subvolume delete command, and then try that first balance command.

But if your volume goes back to 2013, then I think your best strategy is to back up the data, do a factory reset, rebuild the NAS, and restore the data from backup.  Netgear made some adjustments to the on-disk structures after that, and if you rebuild the NAS you should get better performance.

[Sandshark]

---

@flatpack27 wrote:

If I look at the GUI it shows a block of the disk spaced used by snapshots but when you look at the listings on shares on the GUI no snapshots are found. Equally when you select the snapshot option on shares nothing is shown.

Where are you looking for the snapshots?  It sounds like you expect them to show up in the normal share dispaly -- they don't.  The only place they show up is if you go to the Shares page of the GUI and use the gear icon for the share to select "Recover".. There, you can see them and a right-click allows you to delete them.  It's certainly not the most intuitive way to do it.

[flatpack27]

Thanks for that information

I initially deleted the snapshots only on one share, the one that held the 448Gb of files I had deleted recently, but that didn't appear to make any difference. I deleted all snapshots across all shares as the NaS is used as a backup for files held elsewhere and any important data is also replicated to the cloud. Having done that I've now got just over 450Gb of free space.

Until I can assign time to rebuild are there any commands I need to run to try and ensure the snapshots behave properly going forward just in case I do need to make use of them?

Thanks again for your help.

[StephenB]

---

@flatpack27 wrote:

Until I can assign time to rebuild are there any commands I need to run to try and ensure the snapshots behave properly going forward just in case I do need to make use of them?

The "Smart" keep the monthly snapshots forever, which IMO is a really bad idea.

Switch from "smart snapshots" to "custom snapshots".  Then set retention explicitly - perhaps start with 3 months.  I also suggest checking "Only take snapshots with changes".  Personally I have a snapshot schedule set up for daily snapshots (taken at 11 pm).

Then keep an eye on total snapshot use (requires that "quota" be turned on for the volume).  It if gets too high, then adjust the retention downwards for that share.

[flatpack27]

I setup custom snapshots, and quotas, on all the shares and can snapshots in the GUI.

Thanks again for all your help

Cheers

[Sandshark]

While you have probably already discovered this, it takes the NAS some time to reclaim the space previously used by deleted snapshots (or any large deletion), as it does so in the background.  So, if you want to know the effect of snapshot deletion, you need to wait a few minutes.