Re: MySQL NDB Cluster Backup Issue.

On Thursday, May 28, 2015 at 11:01:34 PM UTC+5:30, Jerry Stuckle wrote:
> On 5/28/2015 10:45 AM, Pinal Shah wrote:
> > Hi All,
> >
> > Currently I am working on NDB cluster and I have issue to take backup of NDB cluster.
> >
> > Issues. :
> >
> > 1) Take lot of time to take backup of database using mysqldump.
> > e.g. In innodb engine using mysqldump command it has taken only 2 to 3 min while in NDB cluster it is taking more than 15 min.
> > 2) Try to take backup from management node as START BACKUP ;
> > But got bellow errors:
> >
> > ndb_mgm> start backup
> > Waiting for completed, this may take several minutes
> > Backup failed
> > * 3001: Could not start backup
> > * Too many triggers: Permanent error: Application error
> > ndb_mgm> Node 3: Backup 20150528 started from 2 has been aborted. Error: 4237
> >
> > Could you please help me on above issues and help me on backup & recovery strategies.
> >
> > Regards,
> > Pinal Shah
> >
>
> Not knowing anything about your system and your database, it's
> impossible to tell what the differences are between the two backups.
> There are way too many variables. You need to look into WHY the
> mysqldump is taking so much longer. We can't see it from here.
>
> As for your problem with backing up from the management node - I've
> never seen this, but then I don't back up from the management node. The
> only similar hits I saw in a Google search were from ten years ago, and
> did not look like your problem.
>
> You really need to dig into this more. Michael is correct in that we
> can't help you with this in a newsgroup; there are just way too many
> possibilities.
>
> --
> ==================
> Remove the "x" from my email address
> Jerry Stuckle
> jstucklex_at_attglobal.net
> ==================

Hi Jerry,

All nodes on Amazon EC2 (7.5 GB RAM).

Please check my config.ini file.

[tcp default]
# TCP/IP options:
# portnumber=2202 # This the default; however, you can use any

                  # port that is free for all the hosts in the cluster
                  # Note: It is recommended that you do not specify the port
                  # number at all and simply allow the default value to be used
                  # instead

#BackupMaxWriteSize=1M
#BackupDataBufferSize=16M
#BackupLogBufferSize=4M
#BackupMemory=16M

BackupReportFrequency=10
MemReportFrequency=30
LogLevelStartup=15
LogLevelShutdown=15
LogLevelNodeRestart=15
LogLevelCheckpoint=15
#DiskPageBufferMemory=64M
DiskPageBufferMemory=256M
SharedGlobalMemory=128M
#SharedGlobalMemory=20M
LongMessageBuffer=32M
BatchSizePerLocalScan=512
RedoBuffer=32M
MaxNoOfExecutionThreads=4
StopOnError=false
LockPagesInMainMemory=1
TimeBetweenEpochsTimeout=32000
TimeBetweenWatchdogCheckInitial=60000
TimeBetweenWatchDogCheck=60000
TransactionDeadlockDetectionTimeout=5000 TimeBetweenLocalCheckpoints=20
TimeBetweenGlobalCheckpoints=1000
TimeBetweenEpochs=100
TransactionInactiveTimeout=6000
#TransactionInactiveTimeout=60000
HeartbeatIntervalDbDb=15000
HeartbeatIntervalDbApi=15000
LcpScanProgressTimeout=300

[ndb_mgmd]
hostname=192.168.1.1
datadir=/usr/local/mysql/mysql-cluster
NodeId=1
[ndb_mgmd]
hostname=192.168.1.2
datadir=/usr/local/mysql/mysql-cluster
NodeId=2

[ndbd]
hostname=192.168.1.3
datadir=/usr/local/mysql/data
NodeId=3
[ndbd]
hostname=192.168.1.4
datadir=/usr/local/mysql/data
NodeId=4

[mysqld]
hostname=192.168.1.1
NodeId=5
[mysqld]
hostname=192.168.1.2
NodeId=6
[mysqld]
hostname=192.168.1.1
NodeId=7
[mysqld]
hostname=192.168.1.2
NodeId=8

Regards,
Pinal Shah Received on Fri May 29 2015 - 07:50:09 CEST