YUM Update - Pre Check Activity steps

Posted on

We encountered a situation when running YUM update. Our Server is RHEL 6.4 registered with Satellite in a RHEL 6.6 channel. We have used this channel to upgrade other servers from 6.4 to 6.6 with no issues.

We are unsure exactly what happened, but it appeared that yum updated some packages and then started throwing read only errors, example on /etc. The engineer issued a ctrl c to terminate the yum update operation. We then tried to undo the yum update, and receieved errors, we tired to repair rpm db, received errors, example "error: rpmdbNextIterator: skipping h# 1601 Header V4 RSA/SHA1 signature: BAD, key ID 2874ffbe". After many attempts and after a Case was opened with RedHat, the rpm database was completely hosed which ultimately lead to a reload of the OS which is what RH support recommended.

This server is a HP DL980 G7. The original design uses two HP SmartArray P812 controllers with a software raid for the OS. I suspect that because there were read only errors /etc, the software raid may have caused the issue. The other 6.4 servers where YUM Update succeeded to 6.6, do not have a software raid configuration. (Please don't make remarks about the raid, this was done before I joined the group and was not my decision)

  1. Are there any pre-checks / steps that can be performed before you run YUM Update to ensure the process will work? I know there is this procedure https://access.redhat.com/articles/3763 but was wondering if anyone else had other steps that should be performed or could help identify potential issues.

  2. Should the software raid be dissolved before running YUM Update ?

Thanks in advance

RK

Responses