From: Jonas Oreland Date: December 17 2004 6:32am Subject: Re: ndbd automaticaly shutdown after creating 31 tables List-Archive: http://lists.mysql.com/cluster/1270 Message-Id: <41C27D7A.9030908@mysql.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Hi Alpesh, I think this bug has been solved in 4.1.8 which was released yesterday. Try that version, and if it still does not work, report it as a bug on http://bugs.mysql.com /Jonas Alpesh wrote: > Hello Oliver, > > I removed that arbitrationrank from config file and still i got the same > error. > My nodes configuration is > > CPU : p4 3.0 GHz and 512MB ram > > I am just running the create table statement not inserting any data, then > also it show me error > after successfully creating 31 table.Tables are large about average 10 > columns in each. > > Whether it is a memory problem. > > Becouse on db node error log i got following error: > > Type of error: error > Message: Pointer too large > Fault ID: 2306 > Problem data: DbaccMain.cpp > Object of reference: DBACC (Line: 1176) 0x00000008 > ProgramName: NDB Kernel > ProcessID: 6408 > TraceFile: /usr/local/mysql/mysql-cluster/ndb_3_trace.log.25 > ***EOM*** > > Pls reply me asap. > > - Alpesh > > ----- Original Message ----- > From: "Olivier Kaloudoff" > To: "Alpesh" > Cc: > Sent: Friday, December 10, 2004 2:04 PM > Subject: Re: ndbd automaticaly shutdown after creating 31 tables > > > >>Hi Alpesh, >> >> >> errno 4009 happens when the API node is not connected >>to the DB Nodes. In your case, it seems that you should read how >>the Arbitration process takes place in the cluster, and play with >>the "ArbitrationRank" field for the DB Nodes, after reading this: >> >>http://dev.mysql.com/doc/mysql/en/MySQL_Cluster_API_Definition.html >> >> Until the DB Nodes roles are fixed, you won't be able to >>do useful things with the API node, so first thing is to fix the >>Arbitration problem. If you reload the "show" command fast at >>ndb_mgm prompt, you'll see that the "Master" for the storage nodes >>is moving .. which is not normal, and shows that the nodes are looking >>for a master; >> >>"Problem data: Arbitrator decided to shutdown this node" >> >> Arbitration is a protocol defined to decide which DB Node will >>become the new authoritative Master when some nodes are shut down in >>the cluster. The goal of Arbitration is to avoid "network partitionning", >>ie: having 2 separate sets of datas, served by two sets of mysql nodes, >>which would be a really bad thing for your datas ... >> >> >>Olivier Kaloudoff >>CKR Solutions Open Source >>Mandelieu, France >>http://www.ckr-solutions.com >> >>On Fri, 10 Dec 2004, Alpesh wrote: >> >> >>>Hello Oliver, >>> >>>Afte changing config.ini ( StopOnError: 0 )... >>>The same error i got after creating 31 tables successfully. Error is >>> >>>1005 - Can't create table './test/users_full_backup.frm' (errno: 4009) >>> >>>Now nodes donot shutdown but API status shown as not connected on >>>management server eventhough it is running on the API node. >>> >>>These following errors found on two node error logs. >>> >>>Date/Time: Friday 10 December 2004 - 12:41:11 >>>Type of error: error >>>Message: Arbitrator shutdown >>>Fault ID: 2305 >>>Problem data: Arbitrator decided to shutdown this node >>>Object of reference: QMGR (Line: 3775) 0x00000008 >>>ProgramName: NDB Kernel >>>ProcessID: 5125 >>>TraceFile: /usr/local/mysql/mysql-cluster/ndb_3_trace.log.12 >>>***EOM*** >>> >>>Date/Time: Friday 10 December 2004 - 12:39:11 >>>Type of error: error >>>Message: Pointer too large >>>Fault ID: 2306 >>>Problem data: DbaccMain.cpp >>>Object of reference: DBACC (Line: 1176) 0x00000008 >>>ProgramName: NDB Kernel >>>ProcessID: 4975 >>>TraceFile: /usr/local/mysql/mysql-cluster/ndb_2_trace.log.6 >>> >>>And the same error i got after creating 31 tables successfully. Error is >>> >>>1005 - Can't create table './test/users_full_backup.frm' (errno: 4009) >>> >>> >>>Regards, >>>Alpesh >>> >>>----- Original Message ----- >>>From: "Olivier Kaloudoff" >>>To: "Alpesh" >>>Cc: >>>Sent: Thursday, December 09, 2004 6:29 PM >>>Subject: Re: ndbd automaticaly shutdown after creating 31 tables >>> >>> >>> >>>>Hi Alpesh, >>>> >>>> >>>> you should take a look at the log files for your ndbd daemons >>>>to find out the reason why they left. >>>> >>>> additionnaly, if you may put the following directive in your >>>>config.ini to enable auto restart of the DB nodes after a failure; >>>> >>>>[NDBD DEFAULT] >>>>StopOnError: 0 >>>> >>>> >>>> >>>>Olivier >>>> >>>>On Thu, 9 Dec 2004, Alpesh wrote: >>>> >>>> >>>>>Hello All, >>>>> >>>>>Can some one help...please... >>>>>I created mysql_cluster with two db nodes and one API node.Config.ini > > is > >>>given below... >>> >>>>>I can create about 30 tables and then both the nodes are automaticaly >>> >>>shutdown... >>> >>>>>and also API status shows not connected... >>>>> >>>>>show command displays .... >>>>> >>>>>[ndbd(NDB)] 2 node(s) >>>>>id=2 (not connected, accepting connect from 192.168.2.247) >>>>>id=3 (not connected, accepting connect from 192.168.2.246) >>>>> >>>>>[ndb_mgmd(MGM)] 1 node(s) >>>>>id=1 @192.168.2.247 (Version: 3.5.3) >>>>> >>>>>[mysqld(API)] 1 node(s) >>>>>id=5 (not connected, accepting connect from 192.168.2.248) >>>>> >>>>> >>>>># config.ini" - showing minimal setup consisting of 1 DB node, >>>>># 1 management server, and 3 MySQL servers. >>>>># The empty default sections are not needed, and are shown only for >>> >>>clarity. >>> >>>>># Storage nodes are required to provide a host name but MySQL Servers >>>>># are not. Thus the configuration can be dynamic as to setting up the >>>>># MySQL Servers. >>>>># If you don't know the host name of your machine, use localhost. >>>>># The DataDir parameter also has a default value, but it is > > recommended > >>>to >>> >>>>># set it explicitly. >>>>># NDBD, MYSQLD, and NDB_MGMD are aliases for DB, API, and MGM >>> >>>respectively >>> >>>>># >>>>>[DB DEFAULT] >>>>>NoOfReplicas=2 >>>>>MaxNoOfOrderedIndexes=1000 >>>>>DataDir=/usr/local/mysql/mysql-cluster >>>>> >>>>>[COMPUTER] >>>>>Id:47 >>>>>HostName: 192.168.2.247 >>>>> >>>>>[COMPUTER] >>>>>Id: 48 >>>>>HostName: 192.168.2.248 >>>>> >>>>>[COMPUTER] >>>>>Id: 46 >>>>>HostName: 192.168.2.246 >>>>> >>>>>[MGM] >>>>>Id:1 >>>>>ExecuteOnComputer:47 >>>>>ArbitrationRank: 1 >>>>>DataDir = /usr/local/mysql/mysql-cluster >>>>> >>>>>[DB] >>>>>Id:2 >>>>>ExecuteOnComputer: 47 >>>>>FileSystemPath: /usr/local/mysql/mysql-cluster >>>>> >>>>>[DB] >>>>>Id:3 >>>>>ExecuteOnComputer: 46 >>>>>FileSystemPath: /usr/local/mysql/mysql-cluster >>>>> >>>>>[API] >>>>>Id:5 >>>>>ExecuteOnComputer:48 >>>>> >>>>> >>>>>Regards, >>>>>Alpesh >>>> >>> >>> > > -- Jonas Oreland, Software Engineer MySQL AB, www.mysql.com