new kopanocore fails in autopkgtest in i386 (fails to start at all actually)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kopanocore (Debian) |
Fix Released
|
Unknown
|
|||
kopanocore (Ubuntu) |
Fix Released
|
Undecided
|
Christian Ehrhardt |
Bug Description
Issue in autopkgtest on migration of recent kopanocore:
=> https:/
TL;DR - issue when installing:
Setting up kopano-server (8.3.4-4ubuntu3) ...
Determining localhost credentials from /etc/mysql/
dbconfig-common: writing config to /etc/dbconfig-
Creating config file /etc/dbconfig-
Creating config file /etc/kopano/
checking privileges on database kopanoserver for kopano-
granting access to database kopanoserver for kopano-
verifying access for kopano-
creating database kopanoserver: success.
verifying database kopanoserver exists: success.
dbconfig-common: flushing administrative password
Processing triggers for libc-bin (2.26-0ubuntu2) ...
Processing triggers for systemd (235-2ubuntu3) ...
done.
Check that we have a running server and can create users...
Unable to open Admin session: network error (0x80040115)
The server is not running, or not accessible through "default:".
Using the -v option (possibly multiple times) may give more hints.
I can reproduce the same in autopkgtest via:
$ autopkgtest --apt-upgrade --shell-fail --apt-upgrade --apt-pocket=
Affects i386 AND amd64 when ran in my autopkgtest
In a container the following should be the same:
# enable proposed
cat <<EOF | debconf-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
kopano kopano-
EOF
DEBIAN_
/etc/init.d/mysql start
DEBIAN_
$ kopano-admin -l | grep -qs "SYSTEM.*Kopano"
=> Reproduced in the container
I ran this three ways now:
1. from bionic
2. from bionic-proposed
3. from bionic, but interactive instead of debconf-
=> Only the one in proposed is failing.
Upgrading the working #3 (with manual config) makes it break as well.
Note there is a new mariadb in proposed which could cause this.
The server / service is actually failed in those environments.
root@bionic-
root@bionic-
● kopano-
Loaded: loaded (/lib/systemd/
Active: failed (Result: exit-code) since Tue 2017-11-21 13:09:03 UTC; 18s ago
Docs: man:kopano-
Process: 4319 ExecStart=
Process: 4318 ExecStartPre=
Process: 4317 ExecStartPre=
Process: 4316 ExecStartPre=
Main PID: 4319 (code=exited, status=127)
Nov 21 13:09:03 bionic-
Nov 21 13:09:03 bionic-
Nov 21 13:09:03 bionic-
Nov 21 13:09:03 bionic-
Sometimes in the log there is a message more and I see it when running the server manually:
root@bionic-
/usr/sbin/
I remember seeing something like that in the past.
It was apparmor then and I see now such messages as well:
[2338345.982219] audit: type=1400 audit(151126988
The new version might need that an extension.
Changed in kopanocore (Ubuntu): | |
status: | New → Confirmed |
Changed in kopanocore (Ubuntu): | |
assignee: | nobody → ChristianEhrhardt (paelzer) |
status: | Confirmed → In Progress |
Changed in kopanocore (Debian): | |
status: | Unknown → New |
Changed in kopanocore (Debian): | |
status: | New → Fix Released |
TL;DR - disabling apparmor helps
Details: sbin/kopano- server r,
for the first deny it needs:
/usr/
The next I'm blocked on is: 7.945:2210) : apparmor="DENIED" operation="connect" info="Failed name lookup - disconnected path" error=-13 namespace= "root// lxd-bionic- i386-interactiv e-no-conf_ <var-lib- lxd>" profile= "/usr/sbin/ kopano- server" name="run/ mysqld/ mysqld. sock" pid=25026 comm="kopano- server" requested_mask="wr" denied_mask="wr" fsuid=113 ouid=112
[2339582.514893] audit: type=1400 audit(151127111
For that it already has #include <abstractions/ mysql> but the disconnected path kills that