[
https://issues.apache.org/jira/browse/GEODE-4098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16294234#comment-16294234
]
ASF GitHub Bot commented on GEODE-4098:
---------------------------------------
igorbarc commented on issue #169: GEODE-4098, add mutex timeout
URL: https://github.com/apache/geode-native/pull/169#issuecomment-352270448
@pivotal-jbarrett Thanks, for review, passed now with the the changes
The issue( lock in client) reproduced in geode 1.2.0 release , under high
load from multiple clients, possibly related to buckets moving, rebalancing at
the time, with single-hop enabled for the relevant region
gemfire::ReadGuard::ReadGuard(ACE_RW_Thread_Mutex&)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ReadWriteLock.hpp:33
gemfire::ClientMetadataService::isBucketMarkedForTimeout(char const*,
int)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ClientMetadataService.cpp:1817
6) gemfire::ThinClientPoolDM::getConnectionFromQueueW(GfErrType*,
std::set<gemfire::ServerLocation, std::less<gemfire::ServerLocation>,
std::allocator<gemfire::ServerLocation> >&, bool, gemfire::TcrMessage&, signed
char&, bool&, bool&, gemfire::SharedPtr<gemfire::BucketServerLocation>
const&)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ThinClientPoolDM.hpp:247
7) gemfire::ThinClientPoolDM::sendSyncRequest(gemfire::TcrMessage&,
gemfire::TcrMessage&, bool, bool,
gemfire::SharedPtr<gemfire::BucketServerLocation>
const&)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ThinClientPoolDM.cpp:1400
8) gemfire::ThinClientPoolDM::sendSyncRequest(gemfire::TcrMessage&,
gemfire::TcrMessage&, bool,
bool)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ThinClientPoolDM.cpp:1328
9) gemfire::ThinClientRegion::executeFunction(char const*,
gemfire::SharedPtr<gemfire::Serializable> const&,
gemfire::SharedPtr<gemfire::CacheableVector>, unsigned char,
gemfire::SharedPtr<gemfire::ResultCollector>, int, unsigned
int)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ThinClientRegion.cpp:2955
10) gemfire::ExecutionImpl::execute(char const*, bool, unsigned int, bool,
bool,
bool)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ExecutionImpl.cpp:226
11) gemfire::ExecutionImpl::execute(char const*, bool, unsigned int, bool,
bool)/jphome/jp/ccjp/mb_ccjp/workspace/GEODE_Maven_Build/geode-client-native/src/cppcache/impl/ExecutionImpl.cpp:83
The issue do not seems to be reproducing with the fix
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> Mutex guarding buckets statuses container has no timeout hanldling
> ------------------------------------------------------------------
>
> Key: GEODE-4098
> URL: https://issues.apache.org/jira/browse/GEODE-4098
> Project: Geode
> Issue Type: Bug
> Components: native client
> Reporter: Igor Barchak
> Fix For: 1.2.0
>
>
> Mutex guarding buckets statuses container has no timeout handling,in
> ClientMetadataService.cpp
> Can cause messages stuck on client
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
