On Nov 13, 2012, at 6:28 PM, Piotr Szymaniak wrote: > On Tue, Nov 13, 2012 at 12:19:08AM +0400, fwdnilfsml.to.11df@xxxxxxxx wrote: >> Hello! >> Some time ago I've noticed, that "flush" kernel process associated with >> device running nilfs2 filesystem, generate huge amount of IO activity >> (using 100% CPU time). "iotop" utility show speeds about 5-15 GB\S on >> that "flush" process, but no actual disk activity happens. >> This condition lasts about 1-2 minutes, and then it stops for some time >> (usually 10-20 minutes). > > I can confirm this (but never bothered to check what caused it). But I > saw some huge load (few days ago) on 1 CPU core that was stopped > immediately after runnig "sync". This could be the same as mentioned > here. > What operations were on file system before the issue reproduction? I simply need to understand the way of issue reproduction on my side. > And I just checked that machine and it eats my CPU. ps ax attached. > I see some similarities in these both cases. You mount rootfs with definition the special configuration file also: 2175 ? Ss 0:00 /sbin/nilfs_cleanerd -c /etc/nilfs_cleanerd_rootfs.conf /dev/disk/by-uuid/1aa9e6fb-cf7d-45bd-bbfb-08110a8840b7 So, do you have the reported issue as reproducible in the case of default way of mount or in not rootfs case? With the best regards, Vyacheslav Dubeyko. > > Piotr Szymaniak. > -- > - You know nothing about woman! > - Of course not, I'm a man, so how could I know? > -- Berserk > <ps.out> -- 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
