| Bug #29574 | API Test for 29501 with undo file = (5*1024) cause Ndbd FS inconsistency (lgman) | ||
|---|---|---|---|
| Submitted: | 5 Jul 2007 13:30 | Modified: | 14 Mar 2009 16:06 |
| Reporter: | Jonathan Miller | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Cluster: Disk Data | Severity: | S3 (Non-critical) |
| Version: | 5.1.19 (beta) | OS: | Linux |
| Assigned to: | Jonas Oreland | CPU Architecture: | Any |
[5 Jul 2007 13:33]
Jonathan Miller
cluster log
Attachment: ndb_1_cluster.log (text/x-log), 29.32 KiB.
[5 Jul 2007 13:33]
Jonathan Miller
config file
Attachment: config.ini (application/octet-stream, text), 827 bytes.
[5 Jul 2007 13:33]
Jonathan Miller
error log
Attachment: ndb_3_error.log (text/x-log), 568 bytes.
[5 Jul 2007 13:33]
Jonathan Miller
trace file
Attachment: ndb_3_trace.tar.gz (application/x-gzip, text), 47.66 KiB.
[13 Mar 2009 7:51]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/69094 2867 Jonas Oreland 2009-03-13 ndb - bug#29574 - enfore min-limit on undofiles
[13 Mar 2009 8:16]
Bugs System
Pushed into 5.1.32-ndb-6.2.18 (revid:jonas@mysql.com-20090313075121-17lwkop9g7qgocri) (version source revid:jonas@mysql.com-20090313075121-17lwkop9g7qgocri) (merge vers: 5.1.32-ndb-6.2.18) (pib:6)
[13 Mar 2009 8:17]
Bugs System
Pushed into 5.1.32-ndb-6.3.24 (revid:jonas@mysql.com-20090313075725-906mori1wx0q8hc1) (version source revid:jonas@mysql.com-20090313075725-906mori1wx0q8hc1) (merge vers: 5.1.32-ndb-6.3.24) (pib:6)
[13 Mar 2009 8:18]
Bugs System
Pushed into 5.1.32-ndb-7.0.4 (revid:jonas@mysql.com-20090313081433-xri9mdj8hb2dd6pt) (version source revid:jonas@mysql.com-20090313081433-xri9mdj8hb2dd6pt) (merge vers: 5.1.32-ndb-7.0.4) (pib:6)
[14 Mar 2009 16:06]
Jon Stephens
Documented bugfix in the NDB-6.2.18, 6.3.24, and 7.0.4 changelogs as follows:
When a log file group had an undo log file whose size was too
small, restarting data nodes failed with Read underflow errors.
As a result of this fix, the minimum allowed INTIAL_SIZE for an
undo log file is now 1M (1 megabyte).
Also noted new minimum in "CREATE LOGFILE GROUP Syntax" and "ALTER LOGFILE GROUP Syntax" sections of the Manual.
Description: Setting the udf too low causes a restarting node to fail with Time: Thursday 5 July 2007 - 08:19:27 Status: Ndbd file system error, restart node initial Message: Read underflow (Ndbd file system inconsistency error, please report a bug) Error: 2816 Error data: LGMAN: File system read failed. OS errno: 1000 Error object: LGMAN (Line: 1856) 0x0000000e Program: /usr/local/mysql/libexec/ndbd Pid: 24810 Trace: /usr/local/run/ndb_3_trace.log.1 Version: Version 5.1.19 (beta) lgman 2816: Uint32 tmp = (consumer.m_current_page.m_ptr_i + sz - 2) % sz; How to repeat: int runBug29501(NDBT_Context* ctx, NDBT_Step* step) { NdbRestarter res; NdbDictionary::LogfileGroup lg; lg.setName("DEFAULT-LG"); lg.setUndoBufferSize(8*1024*1024); if (res.getNumDbNodes() < 2) return NDBT_OK; Ndb* pNdb = GETNDB(step); NdbDictionary::Dictionary* pDict = pNdb->getDictionary(); int node = res.getRandomNotMasterNodeId(rand()); res.restartOneDbNode(node, true, true, false); if(pDict->createLogfileGroup(lg) != 0){ g_err << "Failed to create logfilegroup:" << endl << pDict->getNdbError() << endl; return NDBT_FAILED; } NdbDictionary::Undofile uf; uf.setPath("undofile01.dat"); uf.setSize(5*1024); uf.setLogfileGroup("DEFAULT-LG"); if(pDict->createUndofile(uf) != 0){ g_err << "Failed to create undofile:" << endl << pDict->getNdbError() << endl; return NDBT_FAILED; } res.waitNodesNoStart(&node, 1); res.startNodes(&node, 1); if (res.waitClusterStarted()){ g_err << "Node restart failed" << endl << pDict->getNdbError() << endl; return NDBT_FAILED; } if (pDict->dropLogfileGroup(pDict->getLogfileGroup(lg.getName())) != 0){ g_err << "Drop of LFG Failed" << endl << pDict->getNdbError() << endl; return NDBT_FAILED; } return NDBT_OK; } ./testDict -v -n Bug29501 T1