Netezza Performance Server Replication Services 3.0.2.0 release notes

Features

nzdr status user interface updates

• Monitor node status separately using:

  nzdr status node [--details]

• Added a tabular format for node status with the following columns:
  • node-name
  • reachable
  • health
  • replication-status
  • nps-status
• Monitor high level DB status using:

  nzdr status db [--details]

• Added tabular format for DB status which displays whether the database is in sync.

Granular start and stop of replication

• You now have the option to start and stop replication for a specific database (not just at the node level) using:

  nzdr replication [start|stop] --db-name <DB_NAME> [--node-name] <NODE_NAME>

• You can now see more database-specific info in nzdr status.
• To support granular start and stop of replication, added specific replication field attributes like Capture Status and Apply Status.
• Introduced a new replication state in nzdr status to show whether replication has been suspended only for a specific database.

Facility to prune older acknowledged alerts

You can now prune the older acknowledged alerts using:

nzdr alert prune --prune-age <age> --alert_rcs <list of alert types>

New alerts

The system alerts:

• When the NRS storage reaches 85%.
• When NPS goes down.

Ability to set capture and apply flags on Replication nodes

You can set the following flags in local_config.json at each node to customize the nzbackup/nzrestore flags. Following are the default setting:

"c_capture_opts" : "-disableViewCheck
-skipUnchangedViewDefs
-useCxidZonemap 1",
"c_apply_opts" : ""

NRS logs rotation policy

The NRS logs rotation policy is changed to size based rotation instead of daily rotation.

Kafka version upgrade

Upgraded Kafka version to 3.2.3.

Fixed issues

• Fixed the issue with backup happening on primary if database is in standalone state on replica.
• If nzbackup fails partially, replication will now cleanup DIFF backup directory.
• Fixed the issue with the cpds_deploynrs tool not recognizing the -f option to deploy custom certificates.
• The backupmgr can now remove database from DB_LIST.
• Resolved the kafka-client-ssl.properties file corruption issue caused by nzdr add node command.
• Fixed the failure of NRS incremental backup via filesystem if the full backup is taken via non filesystem (EMC).
• Resolved the replication failure issue with database due to backupset ID directory not being created under temp_backup_dir.
• Added nzdr alert help correction for list option.
• backupmgr.log now shows Info Successfully completed capture for DB when there is an error in queuing the incremental backup.

Known issues

1. Errors when deleting databases from replication

   Under a bidirectional replication setup, you might encounter errors when you are running the nzdr delete db command.

   You can ignore this error message.

2. Deployment errors while running cpds_deploynrs with the -i option

   You might encounter this message while running cpds_deploynrs with the -i option.

    mmlsfs: Command failed. Examine previous error messages to determine cause.

   You can ignore this error message.

3. Update and delete database failure on one of the replication systems

   • If nzdr update db failed on one of the replication systems, the nzdr list db output on both systems might be different.

     WORKAROUND 1:

     To revert to the state from before the update, complete the following steps:

     1. Reach out to IBM Support to fix the failure.
     2. Run nzdr update db on the local system again.

   • If nzdr delete db failed on one of the replication systems, the nzdr list db output still displays the database on the system where it failed to delete the database.

     WORKAROUND 2:

     To manually delete the database, complete the following steps:

     1. Log in to the system where the nzdr list db output displays the database to be deleted.

        ssh e2n1
        docker exec -it nrs_replmgmt bash

     2. Remove the database entry from /var/replmgmt/config/global_config.json.
     3. Remove the database from /var/replmgmt/status/status.json.
     4. Prune all kafka records from the database topic:
        1. Find the database topic name:

           /opt/kafka/bin/kafka-topics.sh --list --bootstrap-server localhost:29092

        2. On the list that was generated, find the topic that contains that database name.
        3. Change directories:

           cd /var/replmgmt/kafka/

        4. Add the topic name from the list to prune.config and set offset to -1.
        5. Update prune.config.

           /opt/kafka/bin/kafka-delete-records.sh --bootstrap-server localhost:29092 --offset-json-file prune.config --command-config /var/replmgmt/config/kafka-client-ssl.properties 

        6. • If it is a primary system, log in to the host container and delete the database directory: /nrsbackup/temp_backup_dir.
           • If it is a replica system, log in to the host container and delete the database directory: /nrsbackup/temp_restore_dir.

4. The latency field in the nzdr status --details output does not include the time takes to restore a given database

5. The GPFS filesystem for NRS in the Cloud Pak for Data System might not automount after GPFS service startup or system restart

   WORKAROUND:

   After GPFS or CPDS starts, to check whether GPFS filesystem for NRS is mounted or not, run the following command (ssh to e2n1 node as root).

   >mount | grep gpfs |nrs
   nrs on /opt/ibm/appliance/storage/nrs type gpfs (rw,relatime,seclabel)

   If the output does not contain NRS filesystem, run the following command to manually mount it.

   [root@e2n1 ~]# /usr/lpp/mmfs/bin/mmmount all
   Fri Jul 21 14:15:13 EDT 2023: mmmount: Mounting file systems ...

6. If you run the command nzdr alert prune to pass 0 to prune-age, the following warning message appears

   WARNING: All acknowledged alerts older than 32 days will be removed.
   Are you sure you want to continue (yes/no)? no

   Ignore this warning message.

Limitations

Upgrade paths

NRS 3.0.2.0 can be upgraded directly from the following versions:

• 3.0.1.0
• 3.0.1.1