Closed andriytk closed 5 years ago
The fix for the network-trasport-tcp pkg was landed on GitHub and we take the fixed version since commit a88b4e56.
closed
Actually, the diff is not so huge, see https://github.com/haskell-distributed/network-transport-tcp/pull/86. Let's wait for a few weeks until it will be landed there. If not, we will do it locally here.
Kindly describe — in the merge commit message and in the merge request description — how this chain of commits solves the problem. It's hard to tell looking at the diff alone — the diff is huge.
Would you prefer all 18 commits to be merged as they are, or would you rather cleanup them a bit (e.g., squash "revert" commits)?
In the former case I'll create an additional "merge" commit.
added 1 commit
added 8 commits
added 4 commits
added 4 commits
changed the description
changed the description
changed title from WIP: {-check network-transport change for EndPointAddress Binary instance implementation-} to WIP: {+HALON-876: fix 'Data.Binary.Get.runGet at position NNN: not enough bytes' issue+}
added 4 commits
added 3 commits
added 3 commits
master
added 5 commits
master
[TEST] @dmitriy.chumak this is a notification test, please ignore it
Sometimes during cluster start/stop or when some of the halond processes crashes (due to an SSU crash, especially the RC one on a multi-TS-nodes setup) or when this process restarts later - some of the halond processes may fail with the following error in the system log and get stuck after this: