[Xen-users] ext4_mb_generate_buddy remounting read-only on domU

[Sergi Seira <s.seira@xxxxxxxxx>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

has anyone had ext4 filesystem remounting hours after a live-migration?

We've been having those 'read only remounts' always hours after a 
live-migration.
We are suspecting something is getting corrupted after a live migration, and 
under load an error bumps:

 kernel: [3424760.190005] EXT4-fs error (device xvda3): ext4_mb_generate_buddy: 
EXT4-fs: group 55: 6102 blocks in bitmap, 6101 in gd
 kernel: [3424760.190059] Aborting journal on device xvda3-8.
 kernel: [3424760.191197] EXT4-fs (xvda3): Remounting filesystem read-only
 kernel: [3424760.204396] EXT4-fs (xvda3): delayed block allocation failed for 
inode 374665 at logical offset 318 with max blocks 1 with error -30
 kernel: [3424760.204447] This should not happen!!  Data will be lost
 kernel: [3424760.204483] EXT4-fs (xvda3): ext4_da_writepages: jbd2_start: 1015 
pages, ino 374665; err -30

We have to fsck the volume from dom0 before re-creating domU.

Our system runs on :

 dom0 debian6 2.6.32-5-xen-amd64
 xen-hypervisor 4.0.1-2

DomU are:

 domU paravirtual debian6 2.6.32-5-xen-amd64
 3 iscsi volumes act as vxda's, swap, root and home (400G)
 7 vcpus
 8 GB RAM

Regards,
Sergi
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJOy3DeAAoJEN00VGSIizZ0wLYIAMMo6RCSeevUshwOGUH9t+XT
18lWh/ZZZiG27gnb1Oi/RyR1kjMf8J/cChw0dOBfWbpZPZCIMTsaNS5Y1FhjJKy2
EKHTxp2oB7TMOgPtqnK2610rMzrOmXUT4okw/hfbulUfz16wTgU6FWWGQT0t8CsE
2HuztLtQS361ZShMpcrMcYrqseFyG1nSoSO1v/JES2+FyX0qVYc5IXuxtMWOpd1G
DCU+lWeS4KTT4gJ1OTAbLInZqQrfQtwQyVpiO5nCi0YSzEz0dbsiV5NPEl/twYrf
cXQXoWckMkRXTIeQz1DvShJNIrpb6w7V4ZoVS9aSmmzqrBCWxWUh5LguJcX9FqU=
=dIDX
-----END PGP SIGNATURE-----

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users