Closed mlococo closed 9 years ago
FYI, discussion related to the OpsCenter update to ComboAMI 2.5 is tracked in the private bugtracker under OPSC-2428.
Thanks! This commit has been signed and merged into the dev-2.5 and 2.5 branches.
I see the commit in the other branches, but this PR looks unmerged. I presume your commit-signing workflow involves cherry-picking and pushing rather than merging?
Correct. I could also merge and then submit a basic edit to sign a new commit. Which way would you prefer going forward? I'm open to either.
No preference at this point, just trying to make sure I understand what's going on when I see something "weird-looking" happen. Whenever we get me a signing key and review the current commit-signing workflow (which I'll need quick-tutorial on, I've never done commit-signing before) I'll let you know if I have any feedback at that point.
Thanks for the quick merge, this will let me move forward with the OpsCenter work.
The cassandra limits config baked into the AMI breaks apt-get installs because it's marked as a config file by all the DSE/DSC packages that contain Cassandra, so apt goes into interactive config-diff-resolution mode. This prevents OpsCenter from using ComboAMI 2.5 (which introduced the custom-limits).
This PR moves that limits file to a non-default location when called with the --raidonly flag (which I believe OpsCenter is the only user of) and unblocks OpsCenter upgrading to ComboAMI 2.5.
There's also a larger question about whether the config ComboAMI is using should be pushed into the RPM/DEB packages so all users can benefit.