conntrack log file taking up 40Gb after 1 month

[Sam Stoelinga]

Conntrack log file which became about 40Gb took up the whole root filesystem after running an 6.1 env for a month. Had to delete the conntrack log file and after that kill the process which opened the log file with `lsof | grep "/var" | grep deleted`

I don't have any specifics to reproduce other than run a 8 node environment with 3 controllers in HA using OVS + vlan.

Version: 6.1 nightly build
  build_number: "273"
  build_id: "2015-06-12_05-40-59"

Edit: After inspecting conntrackd.conf I found that following parameter is present LogFile off so there shouldn't be any logfile but there was a logfile. Can anybody check whether /var/log/conntrackd.log is present?

[Matthew Mosesohn (raytrac3r)]

We should just add "/var/log/conntrack*.log" to our list of files to rotate.

[Sam Stoelinga (sammiestoel)]

It seems to me there shouldn't have been a log in the first place after looking at conntrackd.conf but I could be wrong. I didn't understand why the log file was created and became this big, as after restarting the service the log file wasn't recreated. The option LogFile off was present

[Kyrylo Galanov (kgalanov)]

Hi Sam,

Do you still face this issue?
What is the 'tail' of /var/log/conntrackd.log file?

Best regards,
Kyrylo

[Sam Stoelinga (sammiestoel)]

I did not face this issue after restarting conntrackd and the contrackd.log file is no longer present after the restart.