It’s quite rare to have problems with XFS and inodes exhaustion. Mostly because XFS doesn’t have inode limit in a manner known from other filesystems - it’s using some percentage of whole filesystem as a limit and in most distributions it’s 25%. So it’s really huge amount of inodes. But some tools and distributions lowered limit ex. 5% or 10% and there you could have problems more often.
You could check what is you limit by issuing xfs_info with drive and searching for imaxpct value:
xfs_info
root@zombi:~# xfs_info /srv/backup/
metadane=/dev/mapper/slow-backup isize=256 agcount=17, agsize=2621440 blks
= sectsz=512 attr=2
data = bsize=4096 blocks=44564480, imaxpct=25
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096
ascii-ci=0
log =internal bsize=4096 blocks=20480, version=2
= sectsz=512 sunit=0 blks, lazy-count=1
realtime=brak extsz=4096 blocks=0, rtextents=0
In this case I have 25% and it could be changed dynamically with xfs_growfs -m XX where XX is new percentage of volume capacity.
It’s also possible to change imaxpct on creation time by adding option -i maxpct=XX.
Enjoyed?
