-
Notifications
You must be signed in to change notification settings - Fork 1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
master crash in CondVarAny::notify_all #4770
Comments
We should not use connection migrations so it's not clear how the async fiber and connection fiber find themselves on differrent threads. Next step will be to add verbosity logs around connection migrations and see if it matches the crashes. |
Now I see the same on the same thread:
backtrace:
|
Ok, this is the reason for these crashes: https://github.com/dragonflydb/dragonfly/pull/4482/files#diff-b88cbca1b2c9337ab2a37262f9d516b117f4996da25e2ae173e0e129f2b82b20R312 it's indeed a regression from 1.26.x @adiholden and justifies a patch. 1.27.3 here we go! |
thread_queue_backpressure is a global array of per thread QueueBackpressure objects. We referenced these objects incorrectly in 1.27.0-2. Fixes #4770 Signed-off-by: Roman Gershman <[email protected]>
thread_queue_backpressure is a global array of per thread QueueBackpressure objects. We referenced these objects incorrectly in 1.27.0-2. Fixes #4770 Signed-off-by: Roman Gershman <[email protected]>
thread_queue_backpressure is a global array of per thread QueueBackpressure objects. We referenced these objects incorrectly in 1.27.0-2. Fixes #4770 Signed-off-by: Roman Gershman <[email protected]>
the stack trace below.
notify_all
at dragonfly_connection.cc:1640 notifies the connection fiber that is blocked on the pipelining. Both the notifier and the fiber should be on the same thread,however when inspecting the corefile at
WaitQueue::NotifyAll
, i see thatactive->scheduler
andcntx->scheduler
are not the same.The text was updated successfully, but these errors were encountered: