Failure during transaction CLEANUP right after tserver is restarted #613
Comments
yugabyte-ci
pushed a commit
that referenced
this issue
Nov 30, 2018
…EANUP
Summary:
We were failing to check the return code of the function `LookupTablePeerOrRespond` when CLEANUP request is received by tablet service.
This was causing the following FATAL right after restart during software upgrade on a cluster with SecondaryIndex workload.
```#0 yb::tserver::TabletServiceImpl::CheckMemoryPressure<yb::tserver::UpdateTransactionResponsePB> (this=this@entry=0x24c2e00, tablet=tablet@entry=0x0,
resp=resp@entry=0x14d3d410, context=context@entry=0x7f55b1eb5600) at ../../src/yb/tserver/tablet_service.cc:222
#1 0x00007f55d4c8a881 in yb::tserver::TabletServiceImpl::UpdateTransaction (this=this@entry=0x24c2e00, req=req@entry=0x1057aa90, resp=resp@entry=0x14d3d410, context=...)
at ../../src/yb/tserver/tablet_service.cc:431
#2 0x00007f55d273f28a in yb::tserver::TabletServerServiceIf::Handle (this=0x24c2e00, call=...) at src/yb/tserver/tserver_service.service.cc:267
#3 0x00007f55cff0a3ea in yb::rpc::ServicePoolImpl::Handle (this=0x27ca540, incoming=...) at ../../src/yb/rpc/service_pool.cc:214```
Changed LookupTablePeerOrRespond to return complete result using return value.
Test Plan: Update xdc-user-identity and check that is does not crash and workload is stable.
Reviewers: robert, hector, mikhail, kannan
Reviewed By: mikhail, kannan
Subscribers: kannan, ybase
Differential Revision: https://phabricator.dev.yugabyte.com/D5772
mbautin
pushed a commit
that referenced
this issue
Jul 11, 2019
…ed to the earlier commit 566d6d2 Original commit message: ENG-4240: #613: Fix checking of tablet presence during transaction CLEANUP Summary: We were failing to check the return code of the function `LookupTablePeerOrRespond` when CLEANUP request is received by tablet service. This was causing the following FATAL right after restart during software upgrade on a cluster with SecondaryIndex workload. ```#0 yb::tserver::TabletServiceImpl::CheckMemoryPressure<yb::tserver::UpdateTransactionResponsePB> (this=this@entry=0x24c2e00, tablet=tablet@entry=0x0, resp=resp@entry=0x14d3d410, context=context@entry=0x7f55b1eb5600) at ../../src/yb/tserver/tablet_service.cc:222 #1 0x00007f55d4c8a881 in yb::tserver::TabletServiceImpl::UpdateTransaction (this=this@entry=0x24c2e00, req=req@entry=0x1057aa90, resp=resp@entry=0x14d3d410, context=...) at ../../src/yb/tserver/tablet_service.cc:431 #2 0x00007f55d273f28a in yb::tserver::TabletServerServiceIf::Handle (this=0x24c2e00, call=...) at src/yb/tserver/tserver_service.service.cc:267 #3 0x00007f55cff0a3ea in yb::rpc::ServicePoolImpl::Handle (this=0x27ca540, incoming=...) at ../../src/yb/rpc/service_pool.cc:214``` Changed LookupTablePeerOrRespond to return complete result using return value. Test Plan: Update xdc-user-identity and check that is does not crash and workload is stable. Reviewers: robert, hector, mikhail, kannan Reviewed By: mikhail, kannan Subscribers: kannan, ybase Differential Revision: https://phabricator.dev.yugabyte.com/D5772
mbautin
pushed a commit
to mbautin/yugabyte-db
that referenced
this issue
Jul 16, 2019
…ction CLEANUP
Summary:
We were failing to check the return code of the function `LookupTablePeerOrRespond` when CLEANUP request is received by tablet service.
This was causing the following FATAL right after restart during software upgrade on a cluster with SecondaryIndex workload.
```#0 yb::tserver::TabletServiceImpl::CheckMemoryPressure<yb::tserver::UpdateTransactionResponsePB> (this=this@entry=0x24c2e00, tablet=tablet@entry=0x0,
resp=resp@entry=0x14d3d410, context=context@entry=0x7f55b1eb5600) at ../../src/yb/tserver/tablet_service.cc:222
yugabyte#1 0x00007f55d4c8a881 in yb::tserver::TabletServiceImpl::UpdateTransaction (this=this@entry=0x24c2e00, req=req@entry=0x1057aa90, resp=resp@entry=0x14d3d410, context=...)
at ../../src/yb/tserver/tablet_service.cc:431
yugabyte#2 0x00007f55d273f28a in yb::tserver::TabletServerServiceIf::Handle (this=0x24c2e00, call=...) at src/yb/tserver/tserver_service.service.cc:267
yugabyte#3 0x00007f55cff0a3ea in yb::rpc::ServicePoolImpl::Handle (this=0x27ca540, incoming=...) at ../../src/yb/rpc/service_pool.cc:214```
Changed LookupTablePeerOrRespond to return complete result using return value.
Test Plan: Update xdc-user-identity and check that is does not crash and workload is stable.
Reviewers: robert, hector, mikhail, kannan
Reviewed By: mikhail, kannan
Subscribers: kannan, ybase
Differential Revision: https://phabricator.dev.yugabyte.com/D5772
Note:
This commit provides additional functionality that is logically related to
the earlier commit yugabyte@566d6d2
and supersedes the commit yugabyte@63bae60
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We were failing to check the return code of the function LookupTablePeerOrRespond when CLEANUP request is received by tablet service.
This was causing the following FATAL right after restart during software upgrade on a cluster with SecondaryIndex workload.
The text was updated successfully, but these errors were encountered: