Closed bmorgenthaler closed 1 year ago
Tried removing the config directory and starting fresh to adopt a previous backup and it still is giving the same errors in the logs and won't start.
I solved this for me and this was the key for me UpgradeLevelDbEncryption::Unable to read previous universal storage properties., java.io.IOException: IO error: lock /usr/local/crashplan/conf/udb/LOCK: No locks available
My container is mounting it's config and source directory over NFS (backing up a NAS without native docker). I had the defaults from creating the docker volume for a NFS with portainer in the compose. This had been working just fine but with the latest update it broke. After playing around with the NFS options I have the following compose stanza now and it is working. I think the nolock was the key which was counterintuitive.
volumes:
config:
driver_opts:
type: "nfs"
o: "addr=192.168.0.25,rw,noatime,tcp,nolock"
device: ":/docker/crashplan"
The nolock
option disables locking on the server side.
How did you configure the export of the server side ?
Received a notification that CrashPlan hadn't backed up in a day which usually means an update happened and I need to pull the latest image. So I did that but now I cannot start CrashPlan, the web gui display - cannot connect to back-end service, retry. and the docker logs show the following continuously.
Gist of the long service.log.0 file is here: https://gist.github.com/bmorgenthaler/3dad97b33bc7511e939ad03a22f27df0