Version 1.0.21.3 release notes (April 2020)

What's new

Firmware

• Introduced an FSP upgrade.
• Updated DSN to version 8.2.1.6.

  This version includes security fixes and bug fixes. The update also allows to directly upgrade the firmware without first having to downgrade it.

Db2 Support Tools

db_table_restruct

Added a new db_table_restruct script for table optimization. The utility reorganizes tables in the IAS database. The command reloads the data in an ordered manner providing better performance results for queries executed against reorganized tables. The command addresses a common performance issue which is observed after migration of data from one machine to another. For more information, see db_table_restruct

dbpassword

Added a new dbpassword script to store passwords to Db2 systems. The command can be used to store passwords to Db2 systems in an encrypted keystore on a local client machine. For more information, see dbpassword

Note: The dbsql, dbload, db_migrate, and db_migrate_iias support reading passwords from the keystore.

New parameter in db_migrate: ParallelTargetNodes {yes | no}

A new optional parameter is added to the db_migrate command providing support for parallel load of data to multiple target IAS nodes. If the target machine is a multinode system, the script uses all the available nodes to migrate the data. For more information, see db_migrate.

db_build_dictionary script name change

You no longer have to use the .sh extension when running db_build_dictionary. For command help, see db_build_dictionary.

db_get_table_info

The script provides metadata for the provided table(s). For more, see db_get_table_info

db_sort_order

The script provides the percentage of the data sorted on the filesystem with the help of the synopsis table. For more, see db_sort_order

Root lock-down

In IAS Version 1.0.21.3, root lock-down is disabled during upgrade by default. After the upgrade, it is recommended that you change the root password.

High availability improvements

Security improvements:

• Generating HA REST server SSL certificates instead of using static certificates.
• Change to run etcd using stronger cipher suites instead of defaults (includes an etcd upgrade).

Recovery performance improvements:

• Improved Db2 restart time by killing processes running in parallel across all nodes in the system.
• Enabled partial recovery. MLNs are restarted only on nodes with failed MLNs (instead of all MLNs).

Support data collection improvement:

• Introduced new internal APIs to access Db2 transactions and dynamic cache data (from the HA REST API). Used when Platform Manager detects filesystem usage thresholds to help understand the current activity.

Checking for False db activate failures - Ignore all SQL warnings exceptions when re-activating Db2 on recovery.

FCM improved availability

In an MPP environment, the database manager is now able to recover a failed node without having to restart the entire cluster. This reduces the downtime and disruption caused by a node failure.

AFM Disaster recovery

• Added steps to monitor queue status after successful upgrade on both DR systems. For more information, see Upgrading the software in a File system replication environment and Failback to the old primary system.
• Added apdr changeRole for role reversal. For more information, see Failover when primary system crashes.
• Added apdr clean. The feature is used to uninstall the AFMDR service between two system. For more information, see Uninstalling file system replication.

The apdr command enhancements

• Added a apdr diag command. This feature is used to run network diagnostics between the primary and secondary systems. For more, see apdr command and Testing network performance between the primary and secondary systems
• Added a --check-only parameter to the apdr config command. This parameter can be used to run network diagnostics between the systems before configuring DR solution. For more, see apdr command and Setting up file system replication based DR solution on IAS.

Platform management

Introduced SNMP v1 and v2c support, in addition to v3. You can set version when running ap config --set snmp with snmp_version and snmp_community parameters.

If snmp_version is set to 1 or 2, you have to set the following parameters:

• snmp_community
• snmp_ip_address
• snmp_port

If snmp_version is set to 3, you also have to set following parameters:

• snmp_engine_id
• snmp_security_name
• snmp_security_level
• snmp_auth_protocol
• snmp_auth_key
• snmp_enc_protocol
• snmp_enc_key
• snmp_ip_address
• snmp_port

If you switch between versions 3 and 1 or 2c, the configuration parameters are preserved in the system. For more information, see Configuring SNMP trap notifications

FODC dumps

You can now manage FODC dumps via the dbdiag path.

NodeOS

Added a cron script that restarts services if their mount points exceed the threshold. The script resolves the following issues:

• no space error error/job failure.
• Problems with starting containers due to mount failure (can be observed in /var/log/messages.)
• Dangling dm devices with lots of hanging kdmflush and bioset threads which impact system performance (mostly observed on system (i.e. node0101) running DSX.
• The docker cp command failure.

Call Home

Enhanced the Call Home notification email to support CSP (Salesforce). Support requests are now referred to as cases.

Backup and restore

• Version 1.0.21.3 provides support for incremental backup of a schema followed by full restore of the schema or tables within the schema. The feature allows not only concurrent read access to the schema, but also concurrent insert, update, and delete access. A full description of the feature and its limitations are documented here: https://www.ibm.com/support/pages/node/2877819
• db_backup -history now includes sessions column to display the number of sessions used for a TSM/EMC backup. And for EMC restore, it is recommended to use the same number of sessions during restore as well.

Components

Resolved issues

• The following upgrade fixes are introduced in 1.0.21.3:
  • The dashdb container version not identified correctly after the upgrade;
  • Upgrade failing with Failed to reconfigure Docker;
  • GPFS upgrade issue;
  • Upgrade failing with node0101 unreachable.
• Replaced device mapper storage driver with overlay2 storage driver. This fixes any future thinpool issues during upgrades. With upgrading to version 1.0.21.3, DSX reinstall might still be required.
• Fixed an upgrade issue with Docker not starting.
• Fixed an issue with upgrade removing EMC NetWorker configuration. EMC NetWorker package and configuration are now retained after the upgrade.
• Fixed the following issues with static routing:
  • Fixed issue with static routes and nodes losing their fbond/mbond network interfaces.
  • Static routes can now be deleted on every node using the apsetup utility. For more information, see Deleting static routes
• sys_hw_diags network no longer crashes on a 4 HA or above systems.
• db_restore will not be transferring the ownership for system-generated sequences, only for user-defined sequences.
• Resolved the issue with Platform Manager not being able to establish a master.

Known issues

DSX

You have to reinstall DSX, as all DSX containers are gone after upgrade. This is connected to the conversion from the device mapper storage driver to overlay2 storage. The upgrade procedure removes all docker containers when it changes to overlay from device-mapper. The DSX containers are part of the separate DSX installation bundle, not part of the IAS upgrade image.

Workaround:

Back up and restore DSX/ Watson Studio Local as described in https://content-dsxlocal.mybluemix.net/docs/content/SSAS34_current/local/backup.html.

Note: If you want to disable DSX, you have to uninstall it. To uninstall DSX, run ./InstallPackages/utils/uninstall.sh from the directory where install was run. Usually, the install directory is located in /opt/ibm/appliance/storage/platform. If you can't find the directory, contact IBM Support.

Platform Manager reports storage utilization above threshold

The following alert related to fs.sda8 might be opened for every node after upgrading to 1.0.21.3: STORAGE_UTILIZATION | 901: Storage utilization above threshold | sw://fs.sda8/hadomain1.node1 | WARNING |

You can ignore these warnings.

IBM Spectrum Protect backups fail with Kernel issues

When using the IBM Spectrum Protect client for backup and restore, the Kernel issues might occur, which cause the backups to fail.

Workaround:

1. Edit all dsm.sys files in /opt/ibm/appliance/storage/head/tsm/api/node010X-fab, where X is a node number. Add the following lines to all these dsm.sys files:

   tcpbuffsize 512
   tcpwindowsize 1024
   tcpnodelay no

2. Start the backup again.

db_restore

The db_restore command fails for tables defined as external tables. Tables defined as external tables are not assigned a path and are assumed to be empty.

Issues with schema restore

• Alias objects present in a schema are lost after a schema restore. Warnings for aliases that won't be backed up during backup and showing warnings during restore are displayed.
• Schema restore for ownership of objects includes both db2inst1 and original owner.

  Workaround: Revoke db2inst1 as owner of the objects, for those that have two owners after the restore.

dbload -v parameter issue

Instead of displaying additional information about the load session, the -v parameter prints details about the version, similarly to the -V parameter.

EMC NetWorker backup/restore fails

EMC NetWorker backup/restore fails with SQL2062N. An error occurred while accessing the /usr/lib/libnsrdb2.so media. Reason code is 0. This is occurring due to two parameters that are added in the nmda_db2.cfg file on each node.

Workaround:

1. As the root user, edit the /nsr/apps/config/nmda_db2.cfg file and comment out the NSR_LIBNSRDB2_DEBUG_LEVEL=9 and NSR_DEBUG_LEVEL=9 parameters. This must be done on every node in the container.
2. Rerun backup/restore.