On Mon, 7 Mar 2016 19:41, Greg Freemyer wrote:
On Mon, Mar 7, 2016 at 12:01 PM, Anton Aylward wrote:
On 03/07/2016 10:50 AM, Per Jessen wrote:
Greg Freemyer wrote:
I got a journal dump before and after the failure (journalctl -xb > log). About 500 lines of logs added in the post failure log.
I went through and fixed every minor complaint until my server would run smoothly.
Commenting out this line from fstab was the "fix".
=== #/srv_new/portal_backup_container /home/portal_backup/portal_backup ext4 loop 0 0 ===
So something is not quite right about "/srv_new/portal_backup_container". What happens if you try to mount it manually?
The big thing for me that makes this a pretty major bug is that the above line for some reason made my server unusable as a server. Something about it caused systemd to revert to maintenance mode 15 minutes after boot.
Most probably failure to mount that filesystem - I've seen that before.
Quite possibly so. Trying to mount it manually might show what is going on.
On the face of it, it looks like its trying to do do a "bind" mount (q.v. man page) but without the "--bind". After all the "/srv/.." part is not a device. Does it actually exist? Does the destination exist?
Hmm...
The destination does not exist (it used to).
But ignoring that rather big issue, why does a failed mount of a tertiary filesystem cause systemd to crash and burn.
If no one knows, I'll just open this as a bugzilla.
BTW; This isn't a --bind mount. It is a loopback mount. I'm supposed to have a large file at /srv_new/portal_backup_container that itself is a filesystem that I loopback mount.
I know that is strange, but it is needed to overcome an issue with the NFS mount options setup my my cloud provider. Inside the loopback mount I have total control of the mount options.
Oh! how about a workaround: /etc/fstab: (added noauto to options) /srv_new/portal_backup_container /home/portal_backup/portal_backup ext4 noauto,loop 0 0 and for the service that uses this mount (added to the .service file): [Service] ExecPreStart=/bin/sh -c 'test -d /home/portal_backup/portal_backup || mkdir -p /home/portal_backup/portal_backup' ExecPreStart=/bin/sh -c 'grep /home/portal_backup/portal_backup /proc/mounts || mount /home/portal_backup/portal_backup' this two parts should ensure that a) your system boots normally, b) is the mount is not there, no other service is hurt c) the service that NEEDs the mount can do the mount itself. - Yamaban. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org