M0CRT Posted December 10, 2022 Share Posted December 10, 2022 Hi All I've pondered on this for the last 24 hours and am still scratching my head. Cache pool with x2 identical SSDs. Reported available disk space of 1Tb I've several shares set to cache only: System (87.0Gb) Appdata (35.3Gb) cctv-vm (running my CCTV VM) (554Gb) nzbdownload (Deluge downloads et al) (31.8Gb) The rest of my shares do not use cache at all. When I run a calc on space used for these shares and a DU, the space used and remaining does not calculate up. For example, if I calc my cache only shares (using the Unraid GUI on the shares page), the utilise space comes to a total of 708.1Gb. However, my 'free space' on the cache shows 73.0Gb... With my rough math, I feel like I'm missing around 240Gb...Could someone explain what I may be missing here? I've run a balance and scrub on the cache pool (not expecting that this would 'fix' anything) This a block size issue? Any reason why I may be 'loosing' disk space? I've triple checked and don't have something else on the cache beyond the shares / files above. Maybe file permissions in order to 'calc' the disk space? Yours, awaiting enlightenment and putting straight! :-) Thanks Quote Link to comment
JorgeB Posted December 10, 2022 Share Posted December 10, 2022 Du is not reliable with btrfs, note that image type files like vdisks and docker image for example, will grow with time if not regularly unmapped. Quote Link to comment
M0CRT Posted December 10, 2022 Author Share Posted December 10, 2022 1 hour ago, JorgeB said: Du is not reliable with btrfs, note that image type files like vdisks and docker image for example, will grow with time if not regularly unmapped. Thanks for your feedback Jorge. Any reason why the difference in reported free space compared to usage? Quote Link to comment
Kilrah Posted December 10, 2022 Share Posted December 10, 2022 btrfs is complicated https://ohthehugemanatee.org/blog/2019/02/11/btrfs-out-of-space-emergency-response/ Quote Link to comment
M0CRT Posted December 10, 2022 Author Share Posted December 10, 2022 As a follow up. Calc in Unraid shows 708Gb 'occupied' but on summary page, only 70/80Gb Free...Is the BTRS volume potentially reserving some space? Surely this should be reported as approx. 290Gb free? Quote Link to comment
M0CRT Posted December 10, 2022 Author Share Posted December 10, 2022 (edited) 18 minutes ago, Kilrah said: btrfs is complicated https://ohthehugemanatee.org/blog/2019/02/11/btrfs-out-of-space-emergency-response/ I'll take a look :-). Thanks! UPDATE: Great article...thanks Kilrah! Yes...it's not as 'simple' as it may initially seem! :-) Edited December 10, 2022 by M0CRT Quote Link to comment
JorgeB Posted December 10, 2022 Share Posted December 10, 2022 1 minute ago, M0CRT said: Calc in Unraid shows 708Gb 'occupied' but on summary page, only 70/80Gb Free...Is the BTRS volume potentially reserving some space? Surely this should be reported as approx. 290Gb free? Like mentioned image type files will grow in time (and not show the true size they are occupying) if not unmapped, for example for a Window VM see here, another option is to move those files elsewhere then move them back with cp --sparse=always to "resparsify". Quote Link to comment
M0CRT Posted December 10, 2022 Author Share Posted December 10, 2022 27 minutes ago, JorgeB said: Like mentioned image type files will grow in time (and not show the true size they are occupying) if not unmapped, for example for a Window VM see here, another option is to move those files elsewhere then move them back with cp --sparse=always to "resparsify". That's a great idea Jorge. Thanks. Think that's what the issue may be...I've run the balance on the BTRFS and i'm in a good place there. You are right, it's likely to be the 'allocation' of the images associated with my VM. I'll cp them back with 'resparsify' Quote Link to comment
M0CRT Posted December 10, 2022 Author Share Posted December 10, 2022 38 minutes ago, JorgeB said: Like mentioned image type files will grow in time (and not show the true size they are occupying) if not unmapped, for example for a Window VM see here, another option is to move those files elsewhere then move them back with cp --sparse=always to "resparsify". I have checked the 'cctv big disk' and it looks like it may be fully allocated: image: vdisk4.img file format: raw virtual size: 500 GiB (536870912000 bytes) disk size: 500 GiB Will the resparse still help here? i.e. am I not already allocating all the disk? XML of VM...noting the VM is a 'Synology'... <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='writeback'/> <source file='/mnt/user/cctv-vm/Synology/vdisk4.img'/> <target dev='hde' bus='sata'/> <address type='drive' controller='1' bus='0' target='0' unit='4'/> </disk> Quote Link to comment
JorgeB Posted December 10, 2022 Share Posted December 10, 2022 54 minutes ago, M0CRT said: Will the resparse still help here? Yes, if there's free space on the image, but even if there isn't it might still help, because currently it might be using more than 500GiB (despite reporting only 500GiB used) 1 Quote Link to comment
Solution M0CRT Posted December 10, 2022 Author Solution Share Posted December 10, 2022 1 hour ago, JorgeB said: Yes, if there's free space on the image, but even if there isn't it might still help, because currently it might be using more than 500GiB (despite reporting only 500GiB used) Perfect Jorge. Worked like a charm! :-). Thanks 1 Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.