[DRBD-user] drbd-9.1.1

Philipp Reisner philipp.reisner at linbit.com
Fri Mar 26 11:11:56 CET 2021


Debian bullseyes (Debian 11.0) entered 'hard freeze' two weeks ago and we
have customers that want to use DRBD and LINSTOR on that
platform. Unfortunately, I made a stupid mistake a few weeks back and that
broke the kernel compat for kernel 5.10 (which comes with bullseye). So, here
is a release of the drbd-9.1 branch that works on bullseye and also releases
a few fixes we have in the tree.

What fixes?
 Well, did you ever try to build a LINSTOR cluster containing mips64 and
 x86_64 nodes?
 We did, and we found out that it breaks in a small detail in DRBD's size
 negotiation code if the node's architecture has different page
 sizes. Mips64 has a default page size of 16k.
 And some other esoteric things...

Expect the 9.0.29, and another 9.1.x release in about one month from now.

9.1.1 (api:genl2/proto:110-119/transport:17)
 * fix a temporal deadlock you could trigger when you exercise promotion races
   and mix some read-only openers into the test case
 * fix for bitmap-copy operation in a very specific and unlikely case where
   two nodes do a bitmap-based resync due to disk-states
 * fix size negotiation when combining nodes of different CPU architectures
   that have different page sizes
 * fix a very rare race where DRBD reported wrong magic in a header
   packet right after reconnecting
 * fix a case where DRBD ends up reporting unrelated data; it affected
   thinly allocated resources with a diskless node in a recreate from day0
 * changes to socket buffer sizes get applied to established
connections immediately;
   before it was applied after a re-connect
 * add exists events for path objects
 * fix a merge-mistake that broke compatibility with 5.10 kernels


More information about the drbd-user mailing list