memsql-toolbox Release Notes

2020-03-23 version 1.4.4

  • Added computation of a “fingerprint” for each host to avoid adding the same host multiple times under a different name.
  • Added --online flag for memsql-deploy upgrade that runs online upgrade for MemSQL.
  • Fixed a bug where detectCrashStackTraces check displayed warnings if the memsqlStacks.files.json file was absent.
  • Improved error message for memsql-deploy setup-cluster when the path to the source file is incorrect.
  • Added toolboxConfig collector to memsql-report collect that saves the Toolbox’s configuration to a report file.
  • Added a host collector to memsql-report collect that saves each host’s description to a report file.
  • Fixed a bug where memsql-admin enable-high-availability does not request the correct privilege.
  • Added a snapshot retry before upgrade if a database is re-provisioning.
  • Removed extra escape characters in log outputs to make verbose logs more readable.

2020-03-09 version 1.4.3

  • Improved command memsql-deploy upgrade:
    • Added a disk space check and a database snapshot is taken before an upgrade is started.
    • Added cleanup steps in case of an unsuccessful upgrade.
  • Added memsql-admin put-file command that copies a file to the same path on all hosts.
  • Fixed a bug where memsql-admin stop-node --all --force did not stop any nodes if one node was unavailable.
  • Stopped producing messages if no action will be taken by a command.
  • Fixed a bug where memsql-deploy upgrade did not wait for the database to restart after an upgrade is finished.
  • Fixed a bug in memsql-admin enable-high-availability where the command does not work correctly if two leaf nodes on the same host have different hostnames.
  • Added an additional check to memsql-report that ensures all cluster hosts have SSD drives.
  • Added --validate-env option to memsql-report that will validate if MemSQL can be installed.

2020-02-11 Version 1.4.2

  • Improved the prompt message that is displayed when upgrading databases to the new replication modes in MemSQL 7.0.
  • Improved the speed of health checks that are performed on columnstore tables during upgrades.
  • Improved the way memsql-deploy upgrade waits for node recovery.
  • Fixed a bug where memsql-toolbox-config is not able to resolve a state file but returns incorrect error.
  • Fixed a bug where memsql-deploy downloaded the package twice for online upgrade.
  • Fixed the behavior of disk checks before upgrade: always fail upgrade if there is not enough space, show success message on success only.
  • Fixed a bug where toolbox commands did not close the environment after a command failure.
  • Added command-line flag checks to memsqlctl client.
  • Added possibility to set MaxSSHSessionsPerHost parameter with command line (--ssh-max-sessions) or config file.
  • Fixed bug where the memsql-deploy setup-cluster --cluster-file <cluster-file> command returns nodes with Unknown role.

2019-11-25 Version 1.4.1

  • Fixed a bug where all Toolbox commands were not able to pass the --parallelism flag.
  • Fixed a bug where running memsql-deploy upgrade without any flags would perform a rolling restart on the cluster.
  • Fixed a bug where the user would always be prompted for changes to the replication mode of the databases, instead of only being asked when upgrading to MemSQL 7.0.

2019-11-19 Version 1.4.0

  • Updated memsql-deploy to support upgrade to MemSQL 7.0
  • Added TSV based output for some memsql-report collectors.
  • Fixed a bug where memsql-report check fails to check for blocked queries with time 0.
  • Updated memsql-admin update-config --set-global to automatically run on the Master Aggregator.
  • Updated the securityLimits collector to include the output of /etc/security/limits.d and /proc/PID/limits
  • Updated the numactl checker to check if the sum of numa node memory capacity exceeds machine memory capacity.
  • Added new collectors to memsql-report collect:
    • top
    • informationSchemaMvBackupHistory
  • Added new checkers to memsql-report check:
    • orphanTables

2019-10-09 Version 1.3.0

  • Added support for memsql-deploy setup-cluster to accept a YAML file with cluster configurations.
  • Fixed a bug where memsql-report exits with a status of 0 if it fails.
  • Fixed a bug that caused the output of memsql-report collect to be not parsable by memsql-report check
  • Added new collectors to memsql-report collect:
    • uptime
    • shutdownTime
    • rebootTime
  • Added new checkers to memsql-report check:
    • tracelogOOM
    • tracelogOOD
    • explainRebalancePartitionsChecker
    • longRunningQueries
    • detectCrashStackTraces
    • defaultVariables
  • Updated the checkers defaultKeepaliveSettings and defaultHeartbeatSettings to be part of defaultVariables checker

2019-09-03 Version 1.2.3

  • Fixed a bug where the output of memsql-report collect could be missing information.

2019-08-06 Version 1.2.2

  • Added support for rootless memsql-server installs using tar-based packages.
  • Added the --role flag to memsql-admin start-node to start the nodes by role.
  • Added the --role flag to memsql-admin stop-node to stop the nodes by role.
  • Added the Bind Address column to memsql-admin list-nodes output.
  • Updated memsql-admin delete-node to remove leaves and child aggregators from the cluster before deleting them.
  • Updated the sort order for memsql-admin list-nodes to be consistent with other commands.
  • Improved the readability of the License Expiration field in memsql-admin show-license.
  • Fixed a bug that caused the schema collector in memsql-report collect to not collect procedures from memsql-server version 6.7.11 and above.
  • Added new collectors to memsql-report collect:
    • explainRestoreRedundancy
    • explainRebalancePartitions
    • showClusterStatus
  • Added new checkers to memsql-report check:
    • delayedThreadLaunches

2019-04-22 Version 1.1.1

  • Added the --stop flag to memsql-admin delete-node to stop the nodes before deleting them.
  • Added the --host, --port, --user, and --password flags to memsql-admin query for querying a node directly.
  • Added new checkers to memsql-report check:
    • disconnectedReplicationSlaves
    • interpreterMode
    • leafAverageRoundtripLatency
    • runningBackup
    • secondaryDatabases
    • validLicense
  • Updated checkers in memsql-report check:
    • blockedQueries also looks for sleeping threads with open transactions
    • offlineLeaves renamed to leavesNotOnline
  • Added table-based output for some memsql-report collectors.
  • Shortened the filenames inside a cluster report to avoid name length limits in archive files.

2019-03-19 Version 1.1.0

  • Fixed a bug that prevented NOPASSWD sudo from being considered passwordless sudo.
  • Added the memsql-admin query command.
  • Added more properties to memsql-admin list-nodes and memsql-admin describe-node.
  • Removed the swappiness checker. The setting it checked is not required by MemSQL 6.7.
  • Updated the informationSchemaPipelines collector to include information_schema.PIPELINES_CURSORS and information_schema.PIPELINES_FILES.
  • Added new collectors to memsql-report collect:
    • explainOrphanDatabases
    • informationSchemaMvQueries
    • licenseMetadata
  • Added new checkers to memsql-report check:
    • defaultHeartbeatSettings
    • defaultKeepaliveSettings
    • defaultWorkloadManagement
    • defunctProcesses
    • duplicatePartitionDatabase
    • failedCodegen
    • offlineAggregators
    • offlineLeaves
    • pendingDatabases
    • queuedQueries
    • userDatabaseRedundancy
    • versionHashes
  • Updated the report format to include the source paths of copied files.
  • Updated the env commands to include the resolved configFile location.

2019-01-08 Version 1.0.6

  • Updated memsql-deploy cluster-in-a-box to set passwords that are required by memsqlctl create-node starting in MemSQL 6.7.8.
  • memsql-admin create-node now requires the --password flag.
  • Added the --base-install-dir flag to memsql-admin create-node.
  • Updated memsql-report collect --only to also include the memsqlctlInfo collector by default.
  • Added a safety check to memsql-deploy setup-cluster that prevents enabling High Availability with an odd number of leaves.
  • Added the --skip-disk-check flag to memsql-admin enable-high-availability.

2018-12-04 Version 1.0.5

  • Added support for advanced usage of memsql-admin register-node.

2018-12-04 Version 1.0.4 (internal only)

  • Added the attempted command string to error messages due to failed command execution.
  • Added support for running commands as a Linux user that is part of an unnamed group.
  • Fixed a bug in memsql-deploy install that improperly failed the installation if it was unable to add the target user to the memsql Linux group.
  • Added support for advanced memsqlctl configuration to memsql-report collect-local.
  • Reduced spurious errors from memsql-report check when memsqlctl is not installed on the target hosts.
  • Added support for changing the SQL root password on all nodes in the cluster with memsql-admin change-root-password.

2018-11-27 Version 1.0.3

  • Added advanced memsqlctl configuration flags to memsql-toolbox-config register-host.
  • Enabled parallelism for memsql-report collect.
  • Added the report file path to the output of memsql-report collect and memsql-report collect-local.
  • Fixed a bug in memsql-admin confirmation messages that incorrectly implied the use of --password when it was not used.

2018-11-20 Version 1.0.2

  • Fixed bug preventing the use of memsql-admin update-config --all.
  • Added “host” field to memsql-admin describe-node output.
  • Fixed a bug in memsql-deploy list-versions when there are no hosts registered.
  • Changed permissions logic to use memsqlctl configuration to determine the correct user and group for remote commands.

2018-11-09 Version 1.0.1

  • Fixed bug causing Toolbox to assume passwordless sudo if there were cached sudo credentials in the current session.

2018-11-06 Version 1.0.0

  • Initial release of memsql-toolbox.
Was this article useful?