Kresd crashing randomly

Software SW bugs discussion
,
1

Hallo Turris community,

I have a problem with DNS resolver on my Turris.
Every now and then the PC connected to the LAN interface of Turris stops resolving DNS queries and when I login to Turris and try to resolve a query from there, it fails too:

root@turris:~# dig www.debian.org

; <<>> DiG 9.11.5 <<>> www.debian.org
;; global options: +cmd
;; connection timed out; no servers could be reached

In the Turris syslog I see these messages:

root@turris:~# grep -i crash /var/log/messages
2018-12-10 21:47:51 info procd[]: Instance kresd::instance1 s in a crash loop 6 crashes, 232 seconds since last crash
2018-12-11 01:35:41 info procd[]: Instance kresd::instance1 s in a crash loop 6 crashes, 46 seconds since last crash
2018-12-11 16:23:57 info procd[]: Instance kresd::instance1 s in a crash loop 6 crashes, 286 seconds since last crash

I wasn’t able to find more crash information in system logs (please let me know where to look eventually).

Configuration is:

Turris was reset by the three LED reset mode (2 red and one green) into the factory reset state and the initial configuration wizard has been skipped from step 3.
The Atheros 5G wifi card is in Client mode, connected to access point. The 5G wifi card interface is added into the WAN firewall zone.
The AP provides DHCP/DNS information to Turris and Turris runs DHCP/DNS in default configuration for the LAN clients.
My PC is connected via wire to one of the LAN switch ports.
The wired WAN port is not used.
That’s all.

Normally if I toggle the 5G wifi interface off and on using LuCI, the kresd is started and everything works for some time. After a while, DNS resolving starts to fail and I have to manually stop and start the wireless interface or restart resolver or reboot Turris.

It is very annoying and I have no clue how to workaround this.

Please help with any advice.

Thank you,

Best regards,
Tomas Zemek

1 Like
2

I can’t say I have experience with this, but at least for single crashes, one can usually see some information in the log at the moment of the crash, so I’d look for lines around the ones you’ve found. (BTW, logs are kept only in RAM normally, so reboot erases them.)

If there’s nothing useful, verbose logs typically help a lot to determine the cause.

3

I must agree I have the same problem on my Turris Omina. Randomly crashes kresd with the same error message.

Is anybody form Turris resposible for this forum? Original question is 2 weeks old and still without useful answer.

4

Yes, we haven’t received any logs about this from anyone.

5

I enabled verbose logging, got lots of output, but then suddenly the logging stopped and kresd process was gone.
There was nothing abnormal in the verbose output, no error/warning/debug indication of any unexpected state, the kresd process just silently disappeared.

Is there a way I can see if the process was killed by system? It can’t just disappear without any evidence.

I will start investigating the system on daily basis, but I don’t know where to look anymore. Once I get anything out of the box, I’ll post here.

6

I know about abnormal crashes from lingering TTY sockets, as noone can clean them up:

ls -l /tmp/kresd/tty/

Failed assertions should get printed; I can’t remember how e.g. SIGSEGV will manifest in Turris log. In any case, ending of the verbose log might be enough for us to guess what happened.