[ https://issues.apache.org/jira/browse/CASSANDRA-13265?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15923915#comment-15923915 ]
Christian Esken commented on CASSANDRA-13265: --------------------------------------------- bq. This is not quite correct you can't count drainCount as dropped because some of the drained messages may have been sent during iteration. I looked in more detail, and I think this a flaw in the original code "suggested" me to do this: {{drainedMessages.clear()}} is called twice, and one time would be enough. IMO it would be better to only keep the one at the end of the method and also do the drop-counting for the drained messages there. This would also cover a rather exotic case of the {{catch (Exception e)}} in the {{run()}} method. If an Exception is thrown, then there is a danger of nothing being counted. bq. Using a boxed integer bq. You shouldn't need the check for null? >From a brief check, this refers to a similar point. I saw many configuration >options to allow null and followed that route. I am absolutely happy to make >it non-boxed. bq. The right way to do it is create a branch for all the versions where this is going to be fixed. Start at 2.2, merge to 3.0, merge to 3.11, then merge to trunk. At Github? I can do so. But no PR, right? I saw it mentioned that one should not open PR's for Cassandra on Github as they cannot be handled (it's just a mirror). > Expiration in OutboundTcpConnection can block the reader Thread > --------------------------------------------------------------- > > Key: CASSANDRA-13265 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13265 > Project: Cassandra > Issue Type: Bug > Environment: Cassandra 3.0.9 > Java HotSpot(TM) 64-Bit Server VM version 25.112-b15 (Java version > 1.8.0_112-b15) > Linux 3.16 > Reporter: Christian Esken > Assignee: Christian Esken > Fix For: 2.2.x, 3.0.x, 3.11.x, 4.x > > Attachments: cassandra.pb-cache4-dus.2017-02-17-19-36-26.chist.xz, > cassandra.pb-cache4-dus.2017-02-17-19-36-26.td.xz > > > I observed that sometimes a single node in a Cassandra cluster fails to > communicate to the other nodes. This can happen at any time, during peak load > or low load. Restarting that single node from the cluster fixes the issue. > Before going in to details, I want to state that I have analyzed the > situation and am already developing a possible fix. Here is the analysis so > far: > - A Threaddump in this situation showed 324 Threads in the > OutboundTcpConnection class that want to lock the backlog queue for doing > expiration. > - A class histogram shows 262508 instances of > OutboundTcpConnection$QueuedMessage. > What is the effect of it? As soon as the Cassandra node has reached a certain > amount of queued messages, it starts thrashing itself to death. Each of the > Thread fully locks the Queue for reading and writing by calling > iterator.next(), making the situation worse and worse. > - Writing: Only after 262508 locking operation it can progress with actually > writing to the Queue. > - Reading: Is also blocked, as 324 Threads try to do iterator.next(), and > fully lock the Queue > This means: Writing blocks the Queue for reading, and readers might even be > starved which makes the situation even worse. > ----- > The setup is: > - 3-node cluster > - replication factor 2 > - Consistency LOCAL_ONE > - No remote DC's > - high write throughput (100000 INSERT statements per second and more during > peak times). > -- This message was sent by Atlassian JIRA (v6.3.15#6346)