[
https://issues.apache.org/jira/browse/CURATOR-723?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Amr Elazhary updated CURATOR-723:
---------------------------------
Description:
We have a jar application
we are using the below curator-framework-4.3.0 and zookeeper-3.5.7 on the
client servers for such application
!https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
CURATOR-525 There is a race condition in Curator which might lead to fake
SUSPENDED event and ruin CuratorFrameworkImpl inner state - ASF JIRA (State Bug)
!https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
CURATOR-526 Error logged for valid config - "Invalid config event received:
\{properties}" - ASF JIRA (logging Bug)
Above two Bugs Fixed in 5.0.0/5.2.0
We have the below logs on our application (probably a combination of these two
bugs) which ends up on consuming a lot of CPU and a failure after (due to
open/close connections)
{noformat}
2024-10-18T06:35:27.855+11:00 - o.a.c.f.state.ConnectionStateManager - State
change: SUSPENDED
2024-10-18T06:35:27.855+11:00 - - New connection state : SUSPENDED
2024-10-18T06:35:27.855+11:00 - - Connection to server has been suspended..
2024-10-18T06:35:28.638+11:00 - org.apache.curator.ConnectionState -
Negotiated session timeout: 100002024-10-18T06:35:28.639+11:00 -
o.a.c.f.state.ConnectionStateManager - State change: RECONNECTED
2024-10-18T06:35:28.639+11:00 - o.a.c.f.imps.CuratorFrameworkImpl - Clearing
sleep for 0 operations
2024-10-18T06:35:28.639+11:00 - - New connection state : RECONNECTED
2024-10-18T06:35:28.639+11:00 - - Connection to server has been
unsuspended...2024-10-18T06:35:28.640+11:00 -
o.a.c.framework.imps.EnsembleTracker - New config event received:
{}2024-10-18T06:35:28.640+11:00 - - Reconnect worker starting
2024-10-18T06:35:28.640+11:00 - - Attempting to reconnect to the Herd
2024-10-18T06:35:28.640+11:00 - - Waiting for connection to server
2024-10-18T06:35:28.640+11:00 - - Connected to server
2024-10-18T06:35:28.640+11:00 - - Reconnect worker
completed2024-10-18T06:35:28.647+11:00 - o.a.c.framework.imps.EnsembleTracker -
Invalid config event received: {}{noformat}
+*Our Questions*+
1.For the logging Bug "Invalid config event received" , we need to upgrade the
Curator as per https://issues.apache.org/jira/browse/CURATOR-526 please confirm
2. For the reconnections , we are suspecting that it is related to BUG
https://issues.apache.org/jira/browse/CURATOR-525 , however we need to confirm
this (how?), also what is the trigger for this BUG knowing that we have another
jar applications having the same Curator/Zookeeper version and they are not
facing this BUG if our assumption is correct ? please advice the trigger here
and why it is happening on some applications and not happening for others.
also do we have any other solution rather than upgrading the curator version if
this is the root cause?
3. Can we upgrade the curator to version 5.2.0 without upgrading the Zookeper
knowing that we use Zookeeper is zookeeper-3.5.7.
was:
We have a jar application
we are using the below curator-framework-4.3.0 and zookeeper-3.5.7 on the
client servers for such application
!https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
CURATOR-525 There is a race condition in Curator which might lead to fake
SUSPENDED event and ruin CuratorFrameworkImpl inner state - ASF JIRA (State Bug)
!https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
CURATOR-526 Error logged for valid config - "Invalid config event received:
\{properties}" - ASF JIRA (logging Bug)
Above two Bugs Fixed in 5.0.0/5.2.0
We have the below logs on our application (probably a combination of these two
bugs) which ends up on consuming a lot of CPU and a failure after (due to
open/close connections)
{noformat}
2024-10-18T06:35:27.855+11:00 - o.a.c.f.state.ConnectionStateManager - State
change: SUSPENDED
2024-10-18T06:35:27.855+11:00 - au.com.unico.herd.impl.HerdImpl - New
connection state : SUSPENDED
2024-10-18T06:35:27.855+11:00 - au.com.unico.herd.impl.HerdImpl - Connection to
server has been suspended..
2024-10-18T06:35:28.638+11:00 - org.apache.curator.ConnectionState - Negotiated
session timeout: 100002024-10-18T06:35:28.639+11:00 -
o.a.c.f.state.ConnectionStateManager - State change: RECONNECTED
2024-10-18T06:35:28.639+11:00 - o.a.c.f.imps.CuratorFrameworkImpl - Clearing
sleep for 0 operations
2024-10-18T06:35:28.639+11:00 - au.com.unico.herd.impl.HerdImpl - New
connection state : RECONNECTED
2024-10-18T06:35:28.639+11:00 - au.com.unico.herd.impl.HerdImpl - Connection to
server has been unsuspended...2024-10-18T06:35:28.640+11:00 -
o.a.c.framework.imps.EnsembleTracker - New config event received:
{server.2=xhhuj10717:2888:3888:participant,
server.1=xhhuj10716:2888:3888:participant,
server.5=xhwuj10517:2888:3888:participant,
server.4=xhwuj10516:2888:3888:participant,
server.3=xhhuj10721:2888:3888:participant,
version=0}2024-10-18T06:35:28.640+11:00 - au.com.unico.herd.impl.HerdImpl -
Reconnect worker starting
2024-10-18T06:35:28.640+11:00 - au.com.unico.herd.impl.HerdImpl - Attempting to
reconnect to the Herd
2024-10-18T06:35:28.640+11:00 - au.com.unico.herd.impl.HerdImpl - Waiting for
connection to server
2024-10-18T06:35:28.640+11:00 - au.com.unico.herd.impl.HerdImpl - Connected to
server
2024-10-18T06:35:28.640+11:00 - au.com.unico.herd.impl.HerdImpl - Reconnect
worker completed2024-10-18T06:35:28.647+11:00 -
o.a.c.framework.imps.EnsembleTracker - Invalid config event received:
{server.2=xhhuj10717:2888:3888:participant,
server.1=xhhuj10716:2888:3888:participant,
server.5=xhwuj10517:2888:3888:participant,
server.4=xhwuj10516:2888:3888:participant,
server.3=xhhuj10721:2888:3888:participant, version=0} {noformat}
+*Our Questions*+
1.For the logging Bug "Invalid config event received" , we need to upgrade the
Curator as per https://issues.apache.org/jira/browse/CURATOR-526 please confirm
2. For the reconnections , we are suspecting that it is related to BUG
https://issues.apache.org/jira/browse/CURATOR-525 , however we need to confirm
this (how?), also what is the trigger for this BUG knowing that we have another
jar applications having the same Curator/Zookeeper version and they are not
facing this BUG if our assumption is correct ? please advice the trigger here
and why it is happening on some applications and not happening for others.
also do we have any other solution rather than upgrading the curator version if
this is the root cause?
3. Can we upgrade the curator to version 5.2.0 without upgrading the Zookeper
knowing that we use Zookeeper is zookeeper-3.5.7.
> [
> -
>
> Key: CURATOR-723
> URL: https://issues.apache.org/jira/browse/CURATOR-723
> Project: Apache Curator
> Issue Type: Bug
> Components: Framework
> Affects Versions: 4.3.0
> Reporter: Amr Elazhary
> Priority: Major
>
> We have a jar application
> we are using the below curator-framework-4.3.0 and zookeeper-3.5.7 on the
> client servers for such application
> !https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
> CURATOR-525 There is a race condition in Curator which might lead to fake
> SUSPENDED event and ruin CuratorFrameworkImpl inner state - ASF JIRA (State
> Bug)
> !https://issues.apache.org/jira/s/xd97tr/820010/13pdxe5/_/images/fav-jsw.png!
> CURATOR-526 Error logged for valid config - "Invalid config event received:
> \{properties}" - ASF JIRA (logging Bug)
> Above two Bugs Fixed in 5.0.0/5.2.0
>
> We have the below logs on our application (probably a combination of these
> two bugs) which ends up on consuming a lot of CPU and a failure after (due to
> open/close connections)
> {noformat}
> 2024-10-18T06:35:27.855+11:00 - o.a.c.f.state.ConnectionStateManager -
> State change: SUSPENDED
> 2024-10-18T06:35:27.855+11:00 - - New connection state : SUSPENDED
> 2024-10-18T06:35:27.855+11:00 - - Connection to server has been suspended..
> 2024-10-18T06:35:28.638+11:00 - org.apache.curator.ConnectionState -
> Negotiated session timeout: 100002024-10-18T06:35:28.639+11:00 -
> o.a.c.f.state.ConnectionStateManager - State change: RECONNECTED
> 2024-10-18T06:35:28.639+11:00 - o.a.c.f.imps.CuratorFrameworkImpl -
> Clearing sleep for 0 operations
> 2024-10-18T06:35:28.639+11:00 - - New connection state : RECONNECTED
> 2024-10-18T06:35:28.639+11:00 - - Connection to server has been
> unsuspended...2024-10-18T06:35:28.640+11:00 -
> o.a.c.framework.imps.EnsembleTracker - New config event received:
> {}2024-10-18T06:35:28.640+11:00 - - Reconnect worker starting
> 2024-10-18T06:35:28.640+11:00 - - Attempting to reconnect to the Herd
> 2024-10-18T06:35:28.640+11:00 - - Waiting for connection to server
> 2024-10-18T06:35:28.640+11:00 - - Connected to server
> 2024-10-18T06:35:28.640+11:00 - - Reconnect worker
> completed2024-10-18T06:35:28.647+11:00 - o.a.c.framework.imps.EnsembleTracker
> - Invalid config event received: {}{noformat}
>
> +*Our Questions*+
> 1.For the logging Bug "Invalid config event received" , we need to upgrade
> the Curator as per https://issues.apache.org/jira/browse/CURATOR-526 please
> confirm
> 2. For the reconnections , we are suspecting that it is related to BUG
> https://issues.apache.org/jira/browse/CURATOR-525 , however we need to
> confirm this (how?), also what is the trigger for this BUG knowing that we
> have another jar applications having the same Curator/Zookeeper version and
> they are not facing this BUG if our assumption is correct ? please advice the
> trigger here and why it is happening on some applications and not happening
> for others. also do we have any other solution rather than upgrading the
> curator version if this is the root cause?
> 3. Can we upgrade the curator to version 5.2.0 without upgrading the Zookeper
> knowing that we use Zookeeper is zookeeper-3.5.7.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
