hadoop2.3.0cdh5.0.2 Upgrade to cdh5.7.0

After the son on holiday, work this heart really can't hold, in order to be able to enrich the two days spent this difficult, I decided to get a big project ..... PS: Why am I so looking forward to the holiday, lying on the couch like a dead man is really meaningful ...

Of course version: hadoop2.3.0cdh5.0.2

Machine:

	Nn	Dn	Jn	Rm	Nm	Jh	Hmaster	Hregionserver
Mast1	Is	Is	Is	Is	Is		Is	Is
Mast2	Is	Is	Is	Is	Is		Is	Is
Mast3		Is	Is		Is	Is		Is

Target version: hadoop2.6.0cdh5.7.0

Upgrade method: Upgrading unmanaged CDH Using the Command line

Upgrade considerations: ① upgrade from below cdh5.4.0 to cdh5.4.0 or later, HDFs metadata upgrade is required;

② upgrade from less than cdh5.2.0 version: Upgrade HDFs metadata

Upgrade Sentry Database

Upgrading the Hive Database

Upgrading the SQOOP2 Database

③ also make sure that you upgrade the Oozie database and the shared database as follows:

If you uploaded the Spark collection jar file to HDFs, upload the latest version of the file

Upgrade steps:

1. Prepare for the upgrade:

① the Namenode into safe mode, backup Fsimage

[Email protected] conf]$ HDFs haadmin-getservicestate mast1

Active

[[email protected] conf]$ HDFs Dfsadmin-safemode Enter

Safe mode is on

[Email protected] conf]$ HDFs dfsadmin-savenamespace

② confirm that no Hadoop service is running

[email protected] ~]# ps-aef|grep java

Root 9540 8838 0 15:34 pts/0 00:00:00 grep java

③ back up metadata in Namenode (active namenode in ha) (note that if a lock file is found in the process, do it again from the beginning)        

[Email protected] ~]# Cd/app/hdp/dfs/name

[Email protected] name]# Tar-cvf/root/nn_backup_data.tar.

./

./edits/

./edits/current/

./edits/current/edits_inprogress_0000000000000000624

./edits/current/edits_0000000000000000413-0000000000000000533

./edits/current/edits_0000000000000000620-0000000000000000621

./edits/current/edits_0000000000000000618-0000000000000000619

./edits/current/edits_0000000000000000062-0000000000000000180

./edits/current/edits_0000000000000000622-0000000000000000623

./edits/current/edits_0000000000000000038-0000000000000000050

./edits/current/edits_0000000000000000534-0000000000000000615

./edits/current/edits_0000000000000000181-0000000000000000182

./edits/current/edits_0000000000000000284-0000000000000000412

./edits/current/edits_0000000000000000051-0000000000000000061

./edits/current/seen_txid

./edits/current/edits_0000000000000000183-0000000000000000283

./edits/current/version

./edits/current/edits_0000000000000000616-0000000000000000617

./current/

./current/fsimage_0000000000000000000

./current/seen_txid

./current/version

./current/fsimage_0000000000000000000.md5

./namesecondary/

2. Download CDH 5 "1-click" repository

https://archive.cloudera.com/cdh5/one-click-install/redhat/6/x86_64/cloudera-cdh-5-0.x86_64.rpm

3. Update

hadoop2.3.0cdh5.0.2 Upgrade to cdh5.7.0