gluster / project-infrastructure

Issues related to GlusterFs infrastructure components.
0 stars 0 forks source link

builders for centos8_regression don't have passwordless ssh configured #183

Open xhernandez opened 1 year ago

xhernandez commented 1 year ago

Many tests from centos8_regression job fail because the passwordless root ssh seems to be not properly configured in the builders.

From the logs of the jobs:

  16:33:48 ok   5 [     69/    294] <  34> 'gluster --mode=script --wignore volume create secondary replica 2 172.31.9.20:/d/backends/secondary1 172.31.9.20:/d/backends/secondary2 172.31.9.20:/d/backends/secondary3 172.31.9.20:/d/backends/secondary4'
  16:33:48 ok   6 [     14/    615] <  35> 'gluster --mode=script --wignore volume start secondary'
  16:33:48 ok   7 [     15/     38] <  38> 'glusterfs -s 172.31.9.20 --volfile-id primary /mnt/glusterfs/0'
  16:33:48 ok   8 [     43/     64] <  41> 'glusterfs -s 172.31.9.20 --volfile-id secondary /mnt/glusterfs/1'
- 16:33:48 Passwordless ssh login has not been setup with 172.31.9.20 for user root.
  16:33:48 ok   9 [     15/   2626] <  44> 'create_georep_session primary 172.31.9.20::secondary'
xhernandez commented 1 year ago

It seems that the issue only happens on builder-c8-3

mscherer commented 1 year ago

yeah, builder-c8-3 was full, so it is possible it broke something, either during test, or for deployment. Does it still happen after we cleaned the builder ?

mscherer commented 1 year ago

Also, I see that:

# wc -l .ssh/authorized_keys 
313 .ssh/authorized_keys

That's maybe a cause of trouble, as I suspect ssh stop after a while. If the test do not clean after themself, that's a problem :/