[jira] [Commented] (GEODE-10380) When dispatcher socket buffer is full, the ReAuth won't get sent to the client causing a deadlock
[ https://issues.apache.org/jira/browse/GEODE-10380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17554736#comment-17554736 ] ASF subversion and git services commented on GEODE-10380: - Commit 1869f2c06681bb73de727d2080d76c6215db9fb9 in geode's branch refs/heads/support/1.15 from Jinmei Liao [ https://gitbox.apache.org/repos/asf?p=geode.git;h=1869f2c066 ] GEODE-10380: use waitingThreadPool to notify dispatcher at re_auth (#7801) (cherry picked from commit b3fef2a9989ecb5897325a7a84377a8ac7d30028) > When dispatcher socket buffer is full, the ReAuth won't get sent to the > client causing a deadlock > - > > Key: GEODE-10380 > URL: https://issues.apache.org/jira/browse/GEODE-10380 > Project: Geode > Issue Type: Bug > Components: client/server >Affects Versions: 1.15.0 >Reporter: Jinmei Liao >Assignee: Jinmei Liao >Priority: Major > Labels: needsTriage, pull-request-available > Attachments: Screen Shot 2022-06-14 at 5.37.18 PM.png > > > # message dispatcher trying to send a reAuth message to the client, but not > able to because the socket buffer is full. > # client updater is blocked waiting for a regionEntry lock when doing a > remove operation. > # the regionEntry lock is held by a client operation doing destroy on that > same key > # the client operation is waiting for the re_auth op to finish. > # the re_auth command on the server is blocked by the #1 holding the lock -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Commented] (GEODE-10380) When dispatcher socket buffer is full, the ReAuth won't get sent to the client causing a deadlock
[ https://issues.apache.org/jira/browse/GEODE-10380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17554735#comment-17554735 ] ASF subversion and git services commented on GEODE-10380: - Commit b3fef2a9989ecb5897325a7a84377a8ac7d30028 in geode's branch refs/heads/develop from Jinmei Liao [ https://gitbox.apache.org/repos/asf?p=geode.git;h=b3fef2a998 ] GEODE-10380: use waitingThreadPool to notify dispatcher at re_auth (#7801) > When dispatcher socket buffer is full, the ReAuth won't get sent to the > client causing a deadlock > - > > Key: GEODE-10380 > URL: https://issues.apache.org/jira/browse/GEODE-10380 > Project: Geode > Issue Type: Bug > Components: client/server >Affects Versions: 1.15.0 >Reporter: Jinmei Liao >Assignee: Jinmei Liao >Priority: Major > Labels: needsTriage, pull-request-available > Attachments: Screen Shot 2022-06-14 at 5.37.18 PM.png > > > # message dispatcher trying to send a reAuth message to the client, but not > able to because the socket buffer is full. > # client updater is blocked waiting for a regionEntry lock when doing a > remove operation. > # the regionEntry lock is held by a client operation doing destroy on that > same key > # the client operation is waiting for the re_auth op to finish. > # the re_auth command on the server is blocked by the #1 holding the lock -- This message was sent by Atlassian Jira (v8.20.7#820007)
[jira] [Commented] (GEODE-10380) When dispatcher socket buffer is full, the ReAuth won't get sent to the client causing a deadlock
[ https://issues.apache.org/jira/browse/GEODE-10380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17554697#comment-17554697 ] Jinmei Liao commented on GEODE-10380: - PR: https://github.com/apache/geode/pull/7801 > When dispatcher socket buffer is full, the ReAuth won't get sent to the > client causing a deadlock > - > > Key: GEODE-10380 > URL: https://issues.apache.org/jira/browse/GEODE-10380 > Project: Geode > Issue Type: Bug > Components: client/server >Reporter: Jinmei Liao >Assignee: Jinmei Liao >Priority: Major > Labels: needsTriage > Attachments: Screen Shot 2022-06-14 at 5.37.18 PM.png > > > # message dispatcher trying to send a reAuth message to the client, but not > able to because the socket buffer is full. > # client updater is blocked waiting for a regionEntry lock when doing a > remove operation. > # the regionEntry lock is held by a client operation doing destroy on that > same key > # the client operation is waiting for the re_auth op to finish. > # the re_auth command on the server is blocked by the #1 holding the lock -- This message was sent by Atlassian Jira (v8.20.7#820007)