I had a pool of 4 12 TB HDDs in RAIDZ2. Those drives have a real capacity of 10.91 TB after manufacturer rounding, so the total capacity of the pool was around 20 TB.
I added a 5th drive of the same size to the pool, and expected it to increase the total capacity to slightly over 30 TB, but it actually only increased it to 26.31 TB.
Did i do something wrong? Or will the capacity increase over time as the data gets re-packaged across the 5 drives?
There’s actually a fix in the making, but it could be around 1-2 years until it gets available since it’s not a truenas issue, but a openzfs issue. The fix has to land in openzfs first before it lands in truenas.
Okay - i think even with the under reported capacity i have right now, it’ll be several years till i need another drive. so hopefully the fix is in by then.
Just try ALL the “space reporting” tools available to you. If the turenas web gui doesn’t report correctly, open a terminal and try “df -h /[volume]”. There are probably other tools too. Good luck!
And it’s worse than some people think. Not the space reporting… but the difference between zfs rewrite and a re-balancing script.
zfs rewrite works at the block level. So compression, checksum, and other attributes can be changed to whatever is currently selected. While zfs rewrite WILL balance out blocks across multiple vDevs, (if available), to even out space usage. It will NOT redo a file’s data to take advantage of more data columns in a RAID-Zx vDev.
This means a re-balancing script that reads the files and manually copies them, (without block cloning), will reduce parity overhead for these re-balanced files. (In the case of an expanded RAID-Zx vDev…)