| Bug #43699 | ndbmtd crash due to old mgmt server | ||
|---|---|---|---|
| Submitted: | 17 Mar 2009 13:15 | Modified: | 16 Apr 2009 19:05 |
| Reporter: | Guido Ostkamp | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S2 (Serious) |
| Version: | mysql-5.1-telco-6.* | OS: | Solaris |
| Assigned to: | Jonas Oreland | CPU Architecture: | Any |
[18 Mar 2009 7:48]
Sveta Smirnova
Thank you for the report. Which accurate version of old management server do you use?
[18 Mar 2009 8:38]
Guido Ostkamp
Our management server was installed on Feb 26 12:19 from a bazaar built, so most likely it should have been revision id tomas.ulin@sun.com-20090225160230-u4guch19txy3gcew dated Thu 2009-02-26 10:09:51 +0100. Regards Guido Ostkamp
[18 Mar 2009 8:40]
Guido Ostkamp
Sorry, mixed up that date. Correct one is
Wed 2009-02-25 17:02:30 +0100 from
revno: 2585.42.66
revision-id: tomas.ulin@sun.com-20090225160230-u4guch19txy3gcew
parent: frazer@mysql.com-20090225132833-b6amwa2134z392ay
parent: tomas.ulin@sun.com-20090225160116-1uah23jmubh4cevv
committer: Tomas Ulin <tomas.ulin@sun.com>
branch nick: mysql-5.1-telco-6.4
timestamp: Wed 2009-02-25 17:02:30 +0100
message:
merge
[25 Mar 2009 10:35]
Jonathan Miller
Workaround, use new version of mgt srv
[16 Apr 2009 9:13]
Bugs System
Pushed into 5.1.32-ndb-6.2.18 (revid:jonas@mysql.com-20090416090858-wnwzl0s08gwnt1tp) (version source revid:jonas@mysql.com-20090416090858-wnwzl0s08gwnt1tp) (merge vers: 5.1.32-ndb-6.2.18) (pib:6)
[16 Apr 2009 9:14]
Bugs System
Pushed into 5.1.32-ndb-6.3.25 (revid:jonas@mysql.com-20090416091003-yz7qxh53v2z94zbu) (version source revid:jonas@mysql.com-20090416091003-yz7qxh53v2z94zbu) (merge vers: 5.1.32-ndb-6.3.25) (pib:6)
[16 Apr 2009 9:17]
Bugs System
Pushed into 5.1.32-ndb-7.0.6 (revid:jonas@mysql.com-20090416091455-nfi0aqj2phi0urnw) (version source revid:jonas@mysql.com-20090416091114-q9ud1m54v5w47o00) (merge vers: 5.1.32-ndb-7.0.6) (pib:6)
[16 Apr 2009 11:25]
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/72272 2909 Jonas Oreland 2009-04-16 ndb - bug#43699 Fix so we don't try to sprintf("%s", 0)
[16 Apr 2009 19:05]
Jon Stephens
Documented bugfix in the NDB-6.2.18, 6.3.25, and 7.0.6 changelogs as follows:
When trying to use a data node with an older version of the
management server, the data node crashed on startup.
Description: Hello, we are using bzr version frazer@mysql.com-20090317011721-fgtlp5cry4rynf6w dated Tue 2009-03-17 01:17:21 +0000 on Solaris 10 sparc. ndbmtd crashes on startup as follows: t@5 (l@5) terminated by signal SEGV (no mapping at the fault address) 0xffffffff7de3b670: strlen+0x0050: ld [%o2], %o1 Current function is basestring_vsnprintf 59 int ret= vsnprintf(str, size, format, ap); (dbx) where current thread: t@5 [1] strlen(0x0, 0x53, 0x0, 0x0, 0x0, 0x53), at 0xffffffff7de3b670 [2] _ndoprnt(0x1006934bd, 0xffffffff7ccf6fe0, 0xffffffff7dea5fa0, 0x10087a71f, 0x0, 0x1006934bc), at 0xffffffff7dea7d10 [3] vsnprintf(0xffffffff7ccf6eac, 0x0, 0x10069347e, 0xffffffff7ccf6fc0, 0x3dfc2, 0xffffffff7ccf76dc), at 0xffffffff7deaa4ec =>[4] basestring_vsnprintf(str = 0xffffffff7ccf6eac "Connection attempt from management server id=1 with mysql-5.1.32 ndb-6.4.4 incompatible with my\xff|\xcfp0", size = 96U, format = 0x10069347e "Connection attempt from %s id=%d with %s incompatible with %s%s", ap = 0xffffffff7ccf6fc0), line 59 in "basestring_vsnprintf.c" [5] BaseString::vsnprintf(str = 0xffffffff7ccf6eac "Connection attempt from management server id=1 with mysql-5.1.32 ndb-6.4.4 incompatible with my\xff|\xcfp0", size = 96U, format = 0x10069347e "Connection attempt from %s id=%d with %s incompatible with %s%s", ap = 0xffffffff7ccf6fc0), line 468 in "BaseString.cpp" [6] SimulatedBlock::infoEvent(this = 0x197698750, msg = 0x10069347e "Connection attempt from %s id=%d with %s incompatible with %s%s", ...), line 1640 in "SimulatedBlock.cpp" [7] Qmgr::execAPI_REGREQ(this = 0x197698750, signal = 0xffffffff7ccf7700), line 3137 in "QmgrMain.cpp" [8] SimulatedBlock::executeFunction(this = 0x197698750, gsn = 3U, signal = 0xffffffff7ccf7700), line 846 in "SimulatedBlock.hpp" [9] execute_signals(selfptr = 0x1007eb318, q = 0x1007ec388, r = 0x1007ec498, sig = 0xffffffff7ccf7700, max_signals = 253890U, signalIdCounter = 0xffffffff7ccf76dc), line 2429 in "mt.cpp" [10] run_job_buffers(selfptr = 0x1007eb318, sig = 0xffffffff7ccf7700, signalIdCounter = 0xffffffff7ccf76dc), line 2460 in "mt.cpp" [11] mt_job_thread_main(thr_arg = 0x1007eb318), line 2931 in "mt.cpp" [12] ndb_thread_wrapper(_ss = 0x100895840), line 145 in "NdbThread.c" (dbx) frame 7 Current function is Qmgr::execAPI_REGREQ 3137 infoEvent("Connection attempt from %s id=%d with %s " (dbx) print extra extra = (nil) It seems the printf gets illegal parameters in storage/ndb/src/kernel/blocks/qmgr/QmgrMain.cpp" line 3137 if (!compatability_check) { jam(); char buf[NDB_VERSION_STRING_BUF_SZ]; infoEvent("Connection attempt from %s id=%d with %s " "incompatible with %s%s", type == NodeInfo::API ? "api or mysqld" : "management server", apiNodePtr.i, ndbGetVersionString(version, mysql_version, 0, buf, sizeof(buf)), NDB_VERSION_STRING, extra ? extra : 0); apiNodePtr.p->phase = ZAPI_INACTIVE; sendApiRegRef(signal, ref, ApiRegRef::UnsupportedVersion); return; } Regards Guido Ostkamp How to repeat: Use old management server that is incompatible with current ndbmtd version and then try to startup ndbmtd.