Closed Lilja closed 4 years ago
The mount is an SMB share from my NAS. I removed it so it places the data container on the local file system and it works. Perhaps it's best that way? And then perform backup with sqlite export tool to it instead?
@Lilja You need to disable WAL. You can do it by setting an environment variable. See https://github.com/dani-garcia/bitwarden_rs/wiki/Running-without-WAL-enabled
Closed due to inactivity.
Just had the same issue on SWARM cluster using freenas and SMB ;
I fixed it by using the nolock option and mounting the share.
Example :
cat/usr/lib/systemd/system/mnt-smb.mount
[Unit]
Description=SMB mount script
Requires=network.target
After=network.service
[Mount]
What=//amazingNAS.local.io/smb
Where=/mnt/smb
Options=credentials=/tmp/.smbcredentials,nolock,soft,rsize=32768,wsize=32768
Type=cifs
[Install]
WantedBy=multi-user.target
Some people face it here for another tool : https://github.com/Sonarr/Sonarr/issues/1886
However I never had this issue while using CEPH as storage... Might be worth to use it instead of NFS,CIFS and GlusterFS
I had to use "nobrl" flag instead of "nolock" when using the vaultwarden docker image in kubernetes, as "nolock" was being rejected for the CSI Driver for SMB, but it seems to accept "nobrl" instead. Posting for future humans.
The nobrl
is also the correct flag to use for CIFS/SMB to support this.
Though i still suggest to use PostgreSQL or MariaDB/MySQL in these cases.
The
nobrl
is also the correct flag to use for CIFS/SMB to support this. Though i still suggest to use PostgreSQL or MariaDB/MySQL in these cases.
Why? What is the tangible benefit to spin up another database when this is a workable solution?
Also, CIFS is a misnomer. It was a variant name for SMBv1.0 made by Microsoft that never took off, and hasn't been relevant for over 20 years. SMBv2.0/onwards never was and never will be CIFS, it's been SMB the whole time. (I know you're not the only one continuing to use CIFS as a term, but yeah, it's an irrelevant and inaccurate term).
Subject of the issue
thread 'main' panicked at 'Failed to turn on WAL: DatabaseError(__Unknown, "database is locked")'
The container user is able to write to
/data
as I've tried to addcommand: touch /data/hello
.Your environment
Image version 1.13.1
Steps to reproduce
Relevant logs
Full stacktrace: