Verifying the CDP upgrade

This topic lists the steps to verify the CDP upgrade.

Check that the updated CSD is loaded in Cloudera Manager by viewing the Cloudera Manager server log file. In /var/log/cloudera-scm-server/cloudera-scm-server.log, check for the following:
main:com.cloudera.csd.components.CsdRegistryImpl: Installed 1 handlers for CSD [SPECTRUMSCALE_C<CM_VERSION>-<CSD_VERSION>]
where, <CM_VERSION>-<CSD_VERSION> are the updated Cloudera Manager and CSD versions.
For example:
For CDP stack 7.1.7 and CSD version 1.2.0 ensure that SPECTRUMSCALE_C717-1.2.0 is loaded.
```
# grep -ir 'Installed 1 handlers for CSD \[SPECTRUM' /var/log/cloudera-scm-server/cloudera-scm-server.log
2021-08-13 06:05:21,012 INFO main:com.cloudera.csd.components.CsdRegistryImpl: Installed 1 handlers for CSD [SPECTRUMSCALE_C716-1.2.0] 
2021-08-13 06:05:21,034 INFO main:com.cloudera.csd.components.CsdRegistryImpl: Installed 1 handlers for CSD [SPECTRUMSCALE_C717-1.2.0]
```
Verify the HDFS Transparency version on all CES HDFS nodes by running the following command:
```
# rpm -q gpfs.hdfs-protocol
gpfs.hdfs-protocol-3.1.1-5.x86_64
```

Verify that the IBM Storage Scale file system (for example, “gpfs1”) is mounted correctly from one HDFS node:

# mmlsmount gpfs1 -L
File system gpfs1 is mounted on the following four nodes: 
172.16.1.67 c902f10x01 
172.16.1.69 c902f10x02 
172.16.1.73 c902f10x04 
172.16.1.71 c902f10x03

Verify the HDFS status from one HDFS node:

# /usr/lpp/mmfs/hadoop/sbin/mmhdfs hdfs status
c902f10x02.gpfs.net: scaleces: namenode pid is 2267
c902f10x01.gpfs.net: scaleces: namenode pid is 7792 
c902f10x04.gpfs.net: scaleces: datanode pid is 6231 
c902f10x03.gpfs.net: scaleces: datanode pid is 20365

Run mapreduce teragen / terasort job as shown in the following example:

# yarn jar /opt/cloudera/parcels/CDH/lib/hadoop-mapreduce/hadoop-mapreduce-examples.jar teragen 500 test_run
# yarn jar /opt/cloudera/parcels/CDH/lib/hadoop-mapreduce/hadoop-mapreduce-examples.jar terasort test_run test_run_sorted