This seems a legit error, I'm not sure changing what is managing it is going to solve the problem. Sshd fails to write the pid file, nothing should report that is starting correctly.
I was suspecting a selinux error, but audit log doesn't support my theory.
Now I've seen that the error about /var/run/sshd.pid not being readable can be seen also in passing job, butl looking at logstash, first occurrence of both this error and the string containing /var/run/sshd.pid not readable from systemd is 10th January in
This seems a legit error, I'm not sure changing what is managing it is going to solve the problem. Sshd fails to write the pid file, nothing should report that is starting correctly.
I was suspecting a selinux error, but audit log doesn't support my theory.
Now I've seen that the error about /var/run/sshd.pid not being readable can be seen also in passing job, butl looking at logstash, first occurrence of both this error and the string containing /var/run/sshd.pid not readable from systemd is 10th January in
http:// logs.openstack. org/44/ 418344/ 1/check- tripleo/ gate-tripleo- ci-centos- 7-ovb-nonha/ e4e2994/ logs/undercloud /var/log/ undercloud_ install. txt
package diff from a job on 9th Jan doesn't show anything significant.
The openssh package has been changed recently (middle of december) from
openssh- 6.6.1p1- 25.el7. x86_64
to
openssh- 6.6.1p1- 31.el7. x86_64
it may worth a shot opening a bug against openssh-server