0

I'm using docker with device mapper on Redhat 2.6.32. Docker version is 1.5.0.

When docker uses all the data space allocated to docker (Data Space Used value reaches Data Space Total, which is 107G by default) the device mapper file system gets corrupted and I cannot recover unless I delete all docker files and restart docker service, meaning that I lose all my docker containers. The error I get, in that situation, is:

Error response from daemon: Cannot start container mycontainer: Error getting container a40b5c50d48b60df21cebf69b2a266b4d75f8f3e50361fdec1d35f746f1d0d7e from driver devicemapper: Error mounting '/dev/mapper/docker-8:1-2768900-a40b5c50d48b60df21cebf69b2a266b4d75f8f3e50361fdec1d35f746f1d0d7e' on '/DATA/docker/devicemapper/mnt/a40b5c50d48b60df21cebf69b2a266b4d75f8f3e50361fdec1d35f746f1d0d7e': invalid argument

I know we can overcome this by using other copy-on-write file system (btrfs and zfs worked fine) but I am interested in why do we get this strange behavior with device mapper.

[Edit] Kernel version on the machine is 

 Linux 2.6.32-504.16.2.el6.x86_64 #1 SMP Wed Apr 22 06:48:29 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
Emil Salageanu
  • 997
  • 1
  • 10
  • 24
  • Can you provide the complete version of the kernel docker is running ?? – askb Oct 07 '15 at 14:07
  • Yes, I edited the question. – Emil Salageanu Oct 07 '15 at 14:47
  • Looks like you are running an updated kernel as per doc, https://docs.docker.com/v1.5/installation/rhel/ I would also recommend in the mean time raising a support ticket with rh for any known issues – askb Oct 07 '15 at 14:58

0 Answers0