Hi, 14.01.2015, 14:34, Michał Zubkowicz kirjoitti: > I'm using dm-cache on such setup: > dm-cache -> luks -> lvm2 ->ext4 > Kernel version: 3.13.0-43 > > On high disk load (ie when saving virtual machine state) whole system > freezes for a few seconds. > I've tried already to change scheduler (from deadline to cbq), change > ext4 mount options, now i've got: > errors=remount-ro,noatime,nodiratime,discard,barrier=0,relatime > > There is something I should check, or maybe it's a 'normal' behaviour, > or maybe my whole setup is wrong? > Can't really help you directly, but just for the record, I also get whole-system (i.e. not just frozen disks) freezes of 10-60 seconds on heavy disk load on a somewhat similar system: kernel 3.14.27, md-raid6 (16x 4TB) -> lvm with lvm-cache (dm-cache) -> luks -> ext4. Haven't had the time to investigate further, and they are still rather rare for me (~1 per month). They seemed to start when I added dm-cache and luks to the setup. No messages appear in kernel log when it happens. -- Anssi Hannula -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
