From: Date: March 13 2007 11:29am Subject: bk commit into 5.0 tree (jonas:1.2295) BUG#27003 List-Archive: http://lists.mysql.com/commits/21793 X-Bug: 27003 Message-Id: <20070313102918.5CB5D7E20E@perch.ndb.mysql.com> Below is the list of changes that have just been committed into a local 5.0 repository of jonas. When jonas does a push these changes will be propagated to the main repository and, within 24 hours after the push, to the public repository. For information on how to access the public repository see http://dev.mysql.com/doc/mysql/en/installing-source-tree.html ChangeSet@stripped, 2007-03-13 11:29:14+01:00, jonas@stripped +7 -0 ndb - bug#27003 Handle random(not in order) LQHKEYREQ failures during node-restart ndb/src/kernel/blocks/ERROR_codes.txt@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +12 -0 Document new error codes ndb/src/kernel/blocks/dblqh/DblqhMain.cpp@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +16 -3 Handle random(not in order) LQHKEYREQ failures during node-restart ndb/src/kernel/blocks/dbtup/DbtupExecQuery.cpp@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +24 -0 Error codes for various oom problems ndb/src/kernel/blocks/dbtup/DbtupGen.cpp@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +1 -1 move CLEAR_ERROR_INSERT_VALUE to constructor so that it's reasonable to use it for restart testing ndb/src/kernel/blocks/ndbcntr/NdbcntrMain.cpp@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +1 -0 Add error insert for CopyFragRef ndb/test/ndbapi/testNodeRestart.cpp@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +56 -0 Testprg for bug#27003 ndb/test/run-test/daily-basic-tests.txt@stripped, 2007-03-13 11:29:13+01:00, jonas@stripped +4 -0 add testprg # This is a BitKeeper patch. What follows are the unified diffs for the # set of deltas contained in the patch. The rest of the patch, the part # that BitKeeper cares about, is below these diffs. # User: jonas # Host: perch.ndb.mysql.com # Root: /home/jonas/src/50-work --- 1.49/ndb/test/run-test/daily-basic-tests.txt 2007-03-13 11:29:18 +01:00 +++ 1.50/ndb/test/run-test/daily-basic-tests.txt 2007-03-13 11:29:18 +01:00 @@ -425,6 +425,10 @@ cmd: testScan args: -n Bug24447 T1 +max-time: 1000 +cmd: testNodeRestart +args: -n Bug27003 T1 + max-time: 500 cmd: testNodeRestart args: -n Bug15587 T1 --- 1.26/ndb/src/kernel/blocks/ERROR_codes.txt 2007-03-13 11:29:18 +01:00 +++ 1.27/ndb/src/kernel/blocks/ERROR_codes.txt 2007-03-13 11:29:18 +01:00 @@ -489,3 +489,15 @@ 6003 Crash in participant @ CreateTabReq::Prepare 6004 Crash in participant @ CreateTabReq::Commit 6005 Crash in participant @ CreateTabReq::CreateDrop + +TUP: +---- + +4025: Fail all inserts with out of memory +4026: Fail one insert with oom +4027: Fail inserts randomly with oom +4028: Fail one random insert with oom + +NDBCNTR: + +1000: Crash insertion on SystemError::CopyFragRef --- 1.100/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp 2007-03-13 11:29:18 +01:00 +++ 1.101/ndb/src/kernel/blocks/dblqh/DblqhMain.cpp 2007-03-13 11:29:18 +01:00 @@ -9641,6 +9641,15 @@ closeCopyLab(signal); return; }//if + + if (scanptr.p->scanState == ScanRecord::WAIT_LQHKEY_COPY && + scanptr.p->scanErrorCounter) + { + jam(); + closeCopyLab(signal); + return; + } + if (scanptr.p->scanState == ScanRecord::WAIT_LQHKEY_COPY) { jam(); /*---------------------------------------------------------------------------*/ @@ -9717,13 +9726,16 @@ void Dblqh::copyLqhKeyRefLab(Signal* signal) { ndbrequire(tcConnectptr.p->transid[1] == signal->theData[4]); - tcConnectptr.p->copyCountWords -= signal->theData[3]; + Uint32 copyWords = signal->theData[3]; scanptr.i = tcConnectptr.p->tcScanRec; c_scanRecordPool.getPtr(scanptr); scanptr.p->scanErrorCounter++; tcConnectptr.p->errorCode = terrorCode; - closeCopyLab(signal); - return; + + LqhKeyConf* conf = (LqhKeyConf*)signal->getDataPtrSend(); + conf->transId1 = copyWords; + conf->transId2 = tcConnectptr.p->transid[1]; + copyCompletedLab(signal); }//Dblqh::copyLqhKeyRefLab() void Dblqh::closeCopyLab(Signal* signal) @@ -9734,6 +9746,7 @@ // Wait until all of those have arrived until we start the // close process. /*---------------------------------------------------------------------------*/ + scanptr.p->scanState = ScanRecord::WAIT_LQHKEY_COPY; jam(); return; }//if --- 1.22/ndb/src/kernel/blocks/dbtup/DbtupExecQuery.cpp 2007-03-13 11:29:18 +01:00 +++ 1.23/ndb/src/kernel/blocks/dbtup/DbtupExecQuery.cpp 2007-03-13 11:29:18 +01:00 @@ -213,6 +213,30 @@ //--------------------------------------------------- PagePtr pagePtr; Uint32 pageOffset; + + if (ERROR_INSERTED(4025)) + { + signal->theData[0] = 827; + return; + } + if (ERROR_INSERTED(4026)) + { + CLEAR_ERROR_INSERT_VALUE; + signal->theData[0] = 827; + return; + } + if (ERROR_INSERTED(4027) && (rand() % 100) > 25) + { + signal->theData[0] = 827; + return; + } + if (ERROR_INSERTED(4028) && (rand() % 100) > 25) + { + CLEAR_ERROR_INSERT_VALUE; + signal->theData[0] = 827; + return; + } + if (!allocTh(regFragPtr.p, regTabPtr.p, NORMAL_PAGE, --- 1.20/ndb/src/kernel/blocks/dbtup/DbtupGen.cpp 2007-03-13 11:29:18 +01:00 +++ 1.21/ndb/src/kernel/blocks/dbtup/DbtupGen.cpp 2007-03-13 11:29:18 +01:00 @@ -66,6 +66,7 @@ undoPage = 0; totNoOfPagesAllocated = 0; cnoOfAllocatedPages = 0; + CLEAR_ERROR_INSERT_VALUE; // Records with constant sizes }//Dbtup::initData() @@ -570,7 +571,6 @@ switch (startPhase) { case ZSTARTPHASE1: ljam(); - CLEAR_ERROR_INSERT_VALUE; cownref = calcTupBlockRef(0); break; default: --- 1.33/ndb/src/kernel/blocks/ndbcntr/NdbcntrMain.cpp 2007-03-13 11:29:18 +01:00 +++ 1.34/ndb/src/kernel/blocks/ndbcntr/NdbcntrMain.cpp 2007-03-13 11:29:18 +01:00 @@ -180,6 +180,7 @@ break; case SystemError::CopyFragRefError: + CRASH_INSERTION(1000); BaseString::snprintf(buf, sizeof(buf), "Killed by node %d as " "copyfrag failed, error: %u", --- 1.28/ndb/test/ndbapi/testNodeRestart.cpp 2007-03-13 11:29:18 +01:00 +++ 1.29/ndb/test/ndbapi/testNodeRestart.cpp 2007-03-13 11:29:18 +01:00 @@ -1125,6 +1125,59 @@ return NDBT_OK; } +int +runBug27003(NDBT_Context* ctx, NDBT_Step* step) +{ + int result = NDBT_OK; + int loops = ctx->getNumLoops(); + int records = ctx->getNumRecords(); + NdbRestarter res; + + static const int errnos[] = { 4025, 4026, 4027, 4028, 0 }; + + int node = res.getRandomNotMasterNodeId(rand()); + ndbout_c("node: %d", node); + if (res.restartOneDbNode(node, false, true, true)) + return NDBT_FAILED; + + Uint32 pos = 0; + for (Uint32 i = 0; i