On 17 Jul 2019, Peter Grandi said: > I am aware that the opinion that "reboot at all and then check > your filesystems" (and presumably other maintenance operations) > should be or can be performed rarely is very popular among many > sysadmins, being very low cost; to an extreme example I have > seen systems in semi production use for which hardware and > software went EOL in 2007 and untouched since then, and still > "work". But some people get away with it, some don't. The people who believe this include the developers of both the ext4 and xfs filesystems. The recommended ext4 approach is to occasionally snapshot the fs and fsck the snapshot to see if it needs to do anything, and only if does raise the alarm so the sysadmin can schedule a later fsck: the xfs recommendation is only to run a manual xfs_repair iff the filesystem tells you to (and coming down the pipe for xfs is continuous health monitoring and online fsck, hopefully in the end eliminating the need to do xfs_repair ever). Frankly, with modern filesystem sizes, it has become utterly impractical to fsck everything while the system is not fully up and not responding to requests: not unless you *like* your maintenance windows to include multiple hours in which the system is doing nothing other than an fs check that will almost certainly find *nothing whatsoever wrong*. -- NULL && (void)
