Backup restore is not working due to error in Dovecot

jay

Hi,

I just wanted to move a mailcow from one server to another due to lack of storage.

After restore of the backup, I got the following advice:

In most cases it is not required to run a full resync, you can run the command printed below at any time after testing wether the restore process broke a mailbox:

docker exec 4e67799d9c4c doveadm force-resync -A ‘*’

Force a resync now? [y|N] y
Error: auth-master: userdb list: connect(/var/run/dovecot/auth-userdb) failed: Connection refused
Panic: file auth-master.c: line 439 (auth_master_unset_io): assertion failed: (conn->to == NULL)
Error: Raw backtrace: /usr/lib/dovecot/libdovecot.so.0(backtrace_append+0×3d) [0×7feb5789c33d] -> /usr/lib/dovecot/libdovecot.so.0(backtrace_get+0×1e) [0×7feb5789c45e] -> /usr/lib/dovecot/libdovecot.so.0(+0xeca6b) [0×7feb578a5a6b] -> /usr/lib/dovecot/libdovecot.so.0(+0xecaa1) [0×7feb578a5aa1] -> /usr/lib/dovecot/libdovecot.so.0(+0×4ec1e) [0×7feb57807c1e] -> /usr/lib/dovecot/libdovecot.so.0(+0×4c84f) [0×7feb5780584f] -> /usr/lib/dovecot/libdovecot.so.0(+0xd9c49) [0×7feb57892c49] -> /usr/lib/dovecot/libdovecot.so.0(auth_master_user_list_deinit+0xcd) [0×7feb5789467d] -> /usr/lib/dovecot/libdovecot-storage.so.0(mail_storage_service_all_next+0×52) [0×7feb579c5c52] -> /usr/bin/doveadm(+0×31864) [0×562142dc3864] -> /usr/bin/doveadm(doveadm_cmd_ver2_to_mail_cmd_wrapper+0×21a) [0×562142dc44aa] -> /usr/bin/doveadm(doveadm_cmd_run_ver2+0×4df) [0×562142dd49af] -> /usr/bin/doveadm(doveadm_cmd_try_run_ver2+0×37) [0×562142dd4a07] -> /usr/bin/doveadm(main+0×1d1) [0×562142db3e31] -> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xeb) [0×7feb5744009b] -> /usr/bin/doveadm(_start+0×2a) [0×562142db42fa]
c416e1149709

I again tried to do it manually, but got the same error again:

root@mx1:/opt/mailcow-dockerized# docker exec 4e67799d9c4c doveadm force-resync -A ‘*’
Error: auth-master: userdb list: connect(/var/run/dovecot/auth-userdb) failed: Connection refused
Panic: file auth-master.c: line 439 (auth_master_unset_io): assertion failed: (conn->to == NULL)
Error: Raw backtrace: /usr/lib/dovecot/libdovecot.so.0(backtrace_append+0×3d) [0×7fc7ae0de33d] -> /usr/lib/dovecot/libdovecot.so.0(backtrace_get+0×1e) [0×7fc7ae0de45e] -> /usr/lib/dovecot/libdovecot.so.0(+0xeca6b) [0×7fc7ae0e7a6b] -> /usr/lib/dovecot/libdovecot.so.0(+0xecaa1) [0×7fc7ae0e7aa1] -> /usr/lib/dovecot/libdovecot.so.0(+0×4ec1e) [0×7fc7ae049c1e] -> /usr/lib/dovecot/libdovecot.so.0(+0×4c84f) [0×7fc7ae04784f] -> /usr/lib/dovecot/libdovecot.so.0(+0xd9c49) [0×7fc7ae0d4c49] -> /usr/lib/dovecot/libdovecot.so.0(auth_master_user_list_deinit+0xcd) [0×7fc7ae0d667d] -> /usr/lib/dovecot/libdovecot-storage.so.0(mail_storage_service_all_next+0×52) [0×7fc7ae207c52] -> /usr/bin/doveadm(+0×31864) [0×5649914c4864] -> /usr/bin/doveadm(doveadm_cmd_ver2_to_mail_cmd_wrapper+0×21a) [0×5649914c54aa] -> /usr/bin/doveadm(doveadm_cmd_run_ver2+0×4df) [0×5649914d59af] -> /usr/bin/doveadm(doveadm_cmd_try_run_ver2+0×37) [0×5649914d5a07] -> /usr/bin/doveadm(main+0×1d1) [0×5649914b4e31] -> /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xeb) [0×7fc7adc8209b] -> /usr/bin/doveadm(_start+0×2a) [0×5649914b52fa]

Any idea on what to do?

diekuh

Hi, please post your exact steps taken to reproduce it. 🙂

jay

I did a restore as instructed in the documentation.

Installed mailcow on the new server, but didn’t login into the mailcow UI or changed anything.
Updated the mailcow on the old server to the most current version.
Then created a backup using the following command:

BACKUP_LOCATION=/var/backup /opt/mailcow-dockerized/helper-scripts/backup_and_restore.sh backup all

When this was done, I transferred the backup from the old server to the new one, there I started with the restore.

/opt/mailcow-dockerized/helper-scripts/backup_and_restore.sh restore

The script then asked for the path where the backups have been located. After that I was asked which things I want to restore; I pressed 0 to restore everything.

From there on the restore has been made. I can’t provide the logs at the moment since I already closed the SSH session, but the system may generated one in a text file itself? If so, can you tell me where I can find this, so I will paste it here.

Both servers run Ubuntu 18.04 LTS.

Best,

Jay

diekuh

We checked this in an AnyDesk session. I think the problem is, that the script should copy the “old” mailcow.conf to the destination. Afair it does not do that.

jay

Thanks again for a helping hand - so glad to get this awesome support here ❤️🙌