On Thu, Aug 09, 2012 at 01:17:30PM +0200, dexen deVries wrote: > On Thursday 09 of August 2012 12:33:09 you wrote: > > I'm having some issues at uploading the compressed image (it's "just" > > 2.3GB, but with my current connection it would take days, I'll upload > > it on september). > > > > Meanwhile nilfs_cleanerd managed to free some more space (over 1GB), > > although nothing has changed since yesterday, thus everything seems > > back to normality. > > Thanks for the help and see you in september! > > > when i'm in a pinch and want cleanerd to free up some space, i run: > # while true; do killall -HUP nilfs_cleanerd; sleep 4; done > > each `HUP' signal causes cleanerd to both re-read config and run single pass. > > i then monitor progress with > # iostat -dk 1 /dev/sda > > to see if it is moving data around. if not, i lower `protection_period' in > /etc/nilfs_cleanerd.conf; I usually have it set to 7 to 14 days on personal > computers, but when /really/ in a squeeze for free space, i lower it to 24h. > > you may also want to increase `nsegments_per_clean' in nilfs_cleanerd.conf to > 3 (but not much more; each segment causes up to 8MB of read and write upon > each cleanerd pass). Note, you can do all of this using the nilfs-clean command, which will instruct nilfs_cleanerd to do one pass of cleaning. I mostly use: nilfs-clean -p <prot> # Override the protection period in nilfs_cleanerd.conf nilfs-clean -S <segs>/<secs> # Clean <segs> segments every <secs> seconds. I also found the following useful when examining checkpoint usage: lscp | grep -C2 ad The grep captures the active segments at the head of the log, and provides 2 lines of context either side. Christian -- To unsubscribe from this list: send the line "unsubscribe linux-nilfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
