Red Hat

Three RH7 servers patched through a satellite and rebooted Sunday. Two of the three experienced issues with kernel 123.20.1 and were reporting logind and systemd errors, org.freedesktop subscription failure. Users ssh'ing into the two servers had to wait 30 seconds to get the password prompt. Cosign (an Apache mod that allows the use of a remote authentication machine) failed. The third machine was fine, no errors, no problems.

I rolled the two back to 123.13.2 and all of the problems went away. Unfortunately, searching for this on the Internet only brings up similar things from several years ago...nothing current.

Welcome to the world of Systemd. Please always supply your full kernel version number in posts. It should be something like 3.10.0-123.20.1

How did you patch these servers? Using yum update or via individual patches?

It gets more interesting. I first built "test" from scratch last fall. Then I cloned it to "dev." Both of those were running 123.13.2 before the patches, which were done through the Penn State University RH satellite. Both issue the same errors after the update 123.20.1 update and those errors go away when I boot them to 123.13.2. I have several other RH7 servers running 123.20.1 and they have no problems. So I believe this is all unique to this one particular build of dev and test and not indicative of an overall RH patch flaw/fault/bug.

I think it is a good question to ask about how these were updated. In general, when I build a server from scratch, I'll complete the install and then register the machine through the PSU satellite. Ithe recommended procedure is to run a yum update after doing the satellite registration. After that, I will patch using the satellite tools. I don't update with yum after the initial build.