August 11, 20169 yr I have a cache pool with two 500GB samsung SSDs in there and I noticed recently that the pool size has been growing at a rate of about 1GB every 24 hours and I am not sure why. I do not have a share set to use cache except for my cache only shares which contain VM images, one is 215GB and another is 120GB. I also have my docker image on there that says it is 21GB. There are a few other log files in the Docker share but nothing of any significant size. I thought it may have been the metadata growing but it only points to a max of 2GB. I currently have 45GB free and it is shrinking. At the current wait it will fill up and I want to find out why before that happens. Here is the brtfs filesystem df: Data, RAID1: total=463.73GiB, used=422.95GiB System, RAID1: total=32.00MiB, used=80.00KiB Metadata, RAID1: total=2.00GiB, used=669.88MiB GlobalReserve, single: total=224.00MiB, used=0.00B Any help is greatly appreciated. Thanks!
August 12, 20169 yr Author Sure, I have 2 windows 10 VMs with vdisks on the cache, one is 215GB and the other is 129GB (according to browsing the share). I have them both in a vDisks share set for cache only. I also have a Docker share with the docker image set for Cache only. The docker image is 21GB. I have I have a media share set not to use the cache disk, this is where i dump most of my data. I have many sub folders in my media share where the Dockers all point to for settings. I have 7 Dockers consisting of sab, owncloud, deludge, emby, couchpotato, sonarr, and headphones. Let me know if you require any more information. Thanks!
August 12, 20169 yr Install the Krusader docker, navigate to your cache drive contents and go to tools, disk usage. That will tell you what the file system thinks is taking up your space. My suspicion is that you are seeing the VM sparse vdisks grow in usage as windows 10 does its background updates and such. When you allocate vdisks, the size of the file does not immediately use up the free space, it only takes what is actually used. You can put 2 100GB vdisks on a single 120GB volume and not run out of space until the vdisk is actually filled. In your case I suspect you have some files taking up more room than you calculated, but since the only thing currently changing is the vdisks, it seems like your space is going away without cause.
August 12, 20169 yr Author I installed Krusader and took a look at /mnt/cache. I only see my two folders one for vdisks which has two vdisks one reporting 120GB in size an another 200GB. I understand it could be the growing of windows 10 but shouldnt it not grow larger than the sizes I have assigned? The other folder is the Docker folder with my docker image reporting 20GB in size and thats all I see in Krusader. Thanks
August 13, 20169 yr Here is the brtfs filesystem df: Data, RAID1: total=463.73GiB, used=422.95GiB System, RAID1: total=32.00MiB, used=80.00KiB Metadata, RAID1: total=2.00GiB, used=669.88MiB GlobalReserve, single: total=224.00MiB, used=0.00B Any help is greatly appreciated. Thanks! What are these numbers currently? How much change in 2 days?
August 13, 20169 yr Author Here are the current numbers: Data, RAID1: total=463.73GiB, used=423.68GiB System, RAID1: total=32.00MiB, used=80.00KiB Metadata, RAID1: total=2.00GiB, used=676.12MiB GlobalReserve, single: total=240.00MiB, used=0.00B Thanks!
August 13, 20169 yr Author Here is the output: Overall: Device size: 931.52GiB Device allocated: 931.52GiB Device unallocated: 2.11MiB Device missing: 0.00B Used: 848.80GiB Free (estimated): 39.99GiB (min: 39.99GiB) Data ratio: 2.00 Metadata ratio: 2.00 Global reserve: 240.00MiB (used: 0.00B) Data,RAID1: Size:463.73GiB, Used:423.74GiB /dev/sdf1 463.73GiB /dev/sdg1 463.73GiB Metadata,RAID1: Size:2.00GiB, Used:676.30MiB /dev/sdf1 2.00GiB /dev/sdg1 2.00GiB System,RAID1: Size:32.00MiB, Used:80.00KiB /dev/sdf1 32.00MiB /dev/sdg1 32.00MiB Unallocated: /dev/sdf1 1.05MiB /dev/sdg1 1.05MiB
August 15, 20169 yr Author Just got a warning message from unraid that it was filled up to 92% . Krusader again reports no change in size. I am unsure what is filling it up. Thanks
August 15, 20169 yr What is the output of the below command if you ssh to your server? ls -ahl /mnt/cache
August 15, 20169 yr Author Here is the output of the two commands. btrfs fi show: Label: none uuid: af40a6b4-aa87-40ce-aab8-aac9e8b6f0b7 Total devices 2 FS bytes used 426.32GiB devid 1 size 465.76GiB used 465.76GiB path /dev/sdf1 devid 2 size 465.76GiB used 465.76GiB path /dev/sdg1 Label: none uuid: 9242a5b0-b152-40eb-872b-7c78e3de587d Total devices 1 FS bytes used 6.93GiB devid 1 size 20.00GiB used 9.29GiB path /dev/loop0 ls -ahl /mnt/cache: total 16K drwxr-xr-x 1 root root 24 Jun 27 00:58 ./ drwxr-xr-x 10 root root 200 Aug 9 22:21 ../ drwxrwxrwx 1 nobody users 156 Jun 27 01:23 Docker/ drwxrwxrwx 1 nobody users 60 Jun 27 01:50 vDisks/ Thanks!
August 15, 20169 yr btrfs fi df /mnt/cache I think you will probably need to run a btrfs balance operation. I wish I had a live misbehaving BTRFS filesystem to try this on before you do it on your live data, but sadly I don't. I would recommend you make backups of your cache pool contents (minus the docker image file) before you try a balance command. I'm not at all confident of the robustness of BTRFS filesystem, after having a close call with mine several months ago.
August 16, 20169 yr Author Data, RAID1: total=463.73GiB, used=425.78GiB System, RAID1: total=32.00MiB, used=80.00KiB Metadata, RAID1: total=2.00GiB, used=683.12MiB GlobalReserve, single: total=240.00MiB, used=0.00B Should the balance been run with the vms shutdown? Thanks!
August 16, 20169 yr Data, RAID1: total=463.73GiB, used=425.78GiB System, RAID1: total=32.00MiB, used=80.00KiB Metadata, RAID1: total=2.00GiB, used=683.12MiB GlobalReserve, single: total=240.00MiB, used=0.00B Should the balance been run with the vms shutdown? Thanks! Yes, I would make sure there are no open files on the volume before running a balance. At this point I'm grasping at straws, there are stats that you posted that don't look right, and the only info I've found online about it says a balance operation should sort things out. Problem is, I don't have first hand experience with this specific issue to say for sure what's going to fix it. If it were me in this same situation, I'd make a full backup of the cache pool to an array drive, run a byte level verification on the backup, nuke the cache pool, do a fresh format and copy everything back over. I don't trust BTRFS, I had a bad experience with it several months ago. Unfortunately it's the only option available for multiple cache drives.
September 8, 20169 yr Author Just wanted to say I ended up just copying my data off the pool and reformatting both drives and copying the data back to it. That was fine for awhile however it has started to do the same thing again.
Archived
This topic is now archived and is closed to further replies.