When os_patching reports patches but there is not enough space to install them, it reports success

[rnelson0]

Unlike #35, where there is no space to determine if there are patches, this system had enough space to determine there were patches. When the update task ran, yum failed to complete the update but the task reported success. A full log is here, the relevant portion of the debug key is:

Install       7 Package(s)
Upgrade     144 Package(s)
Remove        1 Package(s)

Total download size: 208 M
Downloading Packages:


Error Downloading Packages:
  iproute-2.6.32-57.el6.x86_64: Insufficient space in download directory /var/cache/yum/x86_64/6Server/rhel-6-server-rpms/packages
    * free   0 
    * needed 386 k
  man-pages-overrides-6.10.0-1.el6.noarch: Insufficient space in download directory /var/cache/yum/x86_64/6Server/rhel-6-server-rpms/packages
    * free   0 
    * needed 1.1 M
  ORBit2-2.14.17-7.el6.x86_64: Insufficient space in download directory /var/cache/yum/x86_64/6Server/rhel-6-server-rpms/packages
    * free   0 
    * needed 172 k
  ... <repeate for all packages>

The task also reports "return" : "Success", instead of the expected failure.

[rnelson0]

It looks like run_with_timeout calls yum upgrade -y and only captures the output, and then the Return-Code for the last yum history job is checked, which does not reflect the same job when the disk is full and yum history cannot be update.

Checking the exit code of yum upgrade -y as well as the output could determine if there was an error that may not be reflected in yum history.

[albatrossflavour]

Nice catch, I just asked on the gist : What does a yum history info 9 show? That's what the task uses to determine overall success/failure as the various yum versions are ... unpredictable when it comes to return codes and standard output. That explains it then. I’ll add some extra checking in to determine if yum exited abnormally.

…

On 6 Sep 2018, at 21:16, Rob Nelson wrote: It looks like [run_with_timeout](https://github.com/albatrossflavour/puppet_os_patching/blob/389e257b41fa640a85f2c642690d612e3a3dcd54/tasks/patch_server.rb#L275) calls `yum upgrade -y` and only captures the output, and then the Return-Code for the last [yum history job <id>](https://github.com/albatrossflavour/puppet_os_patching/blob/389e257b41fa640a85f2c642690d612e3a3dcd54/tasks/patch_server.rb#L294) is checked, which does not reflect the same job when the disk is full and yum history cannot be update. Checking the exit code of `yum upgrade -y` as well as the output could determine if there was an error that may not be reflected in `yum history`. -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #39 (comment)

[albatrossflavour]

A fix has been merged into development which adds additional checking to catch this type of issue.