This happens to me too, in at least 50% of the cases (I have tried 4 times so far). Here's my setup:
2x300GB PV 600GB VG 500GB striped LV (2 stripes) lvm 2.02.95-4 linux-image-3.2.0-4-amd64 dmsetup 2:1.02.74-4 Here's what I do: lvcreate -i2 --size 100G --snapshot --name backup /dev/vg1/main # do some work lvremove /dev/vg1/backup and I get the exact error as described above, including frozen I/O on the original volume, /dev/vg1/main). If I do dmsetup resume /dev/mapper/vg1-main the I/O is unfrozen and everything seems to work fine. (Thanks Chris.) There's a lot of I/O activity on the original volume so perhaps that is causing the error. The physical volumes are quite slow. I don't have two snapshots, only one. This is on a replicated production system, but it's mostly reproducable, so perhaps I can try things out if you need me to. Regards, Oskar Liljeblad (os...@osk.mine.nu) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org