Closed lazyfrosch closed 6 years ago
NACK. Patch does not make any sense. Both lines are duplicates of lines that are already there. Also, there is no "8.4" branch. And if there was, it still did not make sense, because 8.4 has no concept of diskless clients.
My mistake, so my problem is actually fixed in master.
But I'm trying to work with 8.9.11
as of now, which updates parts of the shellfunc script. But does not include a default value for status_diskless_client
.
How do you guys handle patches for 8.9.x?
Why would you do that? The regression was introduced 2017-03-22, and fixed 2017-04-26. Don't be scared of numbers. Use the latest drbd-utils release, which is 9.4.0.
I can't just do major updates because of a small problem in a bugfix release.
Currently I'm on 8.9 because it is a very simple setup, without needs for the latest drbd features.
I thought you are still maintaining the 8.9 branch for now.
What are you talking about branches? There has never been a 8.9 branch. We just increased from 8.x.y to 9.x.y, that's it. If you don't want bug fixes, that is your choice, stay at an outdated version with known bugs and backport the fixes you need.
About versioning: 8.9.z was "utils that support 8 and 9". The z was somtimes for bug fixes, sometimes for feature releases, something I did not like, sometimes we had -N for bug fixes only. Then we were at a point where DRBD9 was fully supported, and we jumped to 9.x.y. Now we have a nicer/more standard versioning scheme. .x for for features, .z for bug fixes only. But nothing changed how we develop things, again, there never was a 8.9 "branch"
When using drbd 8.4 the value doesn't get initialized, and never set by
_drbd_set_status_variables_from_events2
so the script always thinks it is on a diskless client.This fix should be ported to the 8.4 branch.