[vitastor] probably corrupted filesystem

[lnsyyj]

Hi @vitalif ,

I have a vistator cluster, I created 11TB of images in vistator, used vitastor-nbd map, formatted XFS to mount it locally, and wrote files to it.

root@vita-1:~# etcdctl --endpoints=192.168.1.215:2379 put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicated","pg_size":3,"pg_minsize":1,"pg_count":256,"failure_domain":"host"}}'

root@vita-1:~# etcdctl --endpoints=http://192.168.1.215:2379/v3 put /vitastor/config/inode/1/1 '{"name":"testimg","size":12094627905536}'

root@vita-1:~# vitastor-nbd map --etcd_address 192.168.1.215:2379/v3 --image testimg

root@vita-1:~# mkfs.xfs /dev/nbd0

root@vita-1:~# mount /dev/nbd0 /home/vita-xfs/

During data writing the vistator-nbd process is gone, when I mount it again,

root@vita-1:~# mount /dev/nbd0 /home/vita-xfs/
mount: /home/vita-xfs: cannot mount; probably corrupted filesystem on /dev/nbd0.

[vitalif]

Can you run vitastor-nbd in foreground (with --foreground 1) and reproduce the crash? Does vitastor-nbd crash again when you try to mount xfs the second time?