[BlueOnyx:22664] Re: Started Crash recovery kernel arming

[Michael Stauber]

Hi all,

> I solved it by killing yum quick enough after boot, removing
> blueonyx-le-acme package, disabling the line in hotfixes.sh that tries
> to repeatedly install it temporarily, yum clean all, yum update and then
> installing blueonyx-le-acme manually again. Either one of yum update or
> yum install blueonyx-le-acme installs a new hotfix.sh which doesn't seem
> to cause the described problems.

Yeah, there was unfortunately a problem with the "swatch" RPM that I had
by accident published for 5209R two days ago. It was supposed to go to
the "BlueOnyx-Testing" YUM repository, but was released to the regular
repo instead.

I realized that mistake within the next 20 minutes and removed it from
the repo, rebuilt the indexes and manually synced all mirrors that I
have direct access to. However, due to the unfortunate timing of the
incident servers in Europe that didn't operate from a cached instance of
the YUM repository data might have pulled the bad "swatch".

If you're affected, do this:

1.) Check with "ps axf|grep yum" if YUM is running and if hotfixes is
running: "ps axf|grep hotfixes". If so, kill all related processes off.

2.) Remove the executable bits from /usr/sausalito/sbin/hotfixes.sh by
running:

chmod 600 /usr/sausalito/sbin/hotfixes.sh

3.) For good measure rebuild your RPM database to get rid of any
inconsistencies or remaining RPM database locks that this might have caused:

rpm --rebuilddb

4.) Do a "yum clean all; yum update -y" and it'll install a fixed
"swatch" RPM. The latest swatch for 5209R is this:

~]# rpm -q swatch
swatch-4.0.0-0BX52.el7.noarch


-- 
With best regards

Michael Stauber