Navigation
This version of the documentation is archived and no longer supported. To learn how to upgrade your version of MongoDB Ops Manager, refer to the upgrade documentation.
You were redirected from a different version of the documentation. Click here to go back.

Ops Manager Server Changelog

Ops Manager Server 3.4

Ops Manager Server 3.4.15

Released on 2018-08-02

  • Update Ops Manager JDK (x86_64) to 8u181.

Ops Manager Server 3.4.14

Released on 2018-05-03

  • Fix: At the completion of an automated restore of a sharded cluster, Ops Manager ensures that the balancer is enabled.
  • Fix: Network charts now display physical bytes in/out instead of logical bytes in/out. That is, if network compression is enabled for MongoDB, the network charts shows the compressed bytes in/out.
  • Upgrade: Bundled JDK to 8u172.
  • Agent Upgrades: Automation Agent 3.2.18.4938-1

Ops Manager Server 3.4.13

Released on 2018-04-05

Ops Manager Server 3.4.12

Released on 2018-02-01

  • The bundled JDK is version 8u162
  • Upgrade jackson-databind to 2.9.3

Ops Manager Server 3.4.11

Released on 2017-12-20

  • Fix: Issue importing a process into Automation running on a custom build.
  • Fix: Explicitly specify the collation {locale: simple} in createIndex during the initial sync of a backup.
  • Fix: Allow clusterAuthMode to be configured when deployment does not use MONGODB-X509 authentication.
  • Fix: Allow importing a process into Automation using the allowConnectionsWithoutCertificates parameter.
  • Security enhancements.

Ops Manager Server 3.4.10

Released on 2017-11-02

  • Agent Upgrades: Automation Agent 3.2.16.2263
  • Fix: credentialstool and encryptiontool load dependencies correctly.
  • Fix: Backup groom jobs no longer get stuck on partially-deleted snapshots.
  • Fix: Some of the Cloud Manager interface no longer shown when creating groups.
  • Fix: where sorting during deletion of backups could run out of memory.
  • Reduce memory usage of backup snapshot deletion.
  • Remove a restriction where a shard key range could not have an equal minimum and maximum.
  • Update JDK to 8u152.

Ops Manager Server 3.4.9

Released on 2017-10-05

  • Agent Upgrades: Automation Agent 3.2.15.2257, Backup Agent 5.0.11.663
  • Fix: Allow the mongos autoSplit configuration parameter to be removed simultaneously with a major version upgrade from MongoDB 3.2 to 3.4.
  • Fix: During the initial sync of a new backup, documents with a compound _id that are moved on disk at the source during the initial sync may be omitted from the backup under rare conditions. This issue affects only MMAPv1.
  • Increase the expiration period for snapshots downloaded during an automated restore and make the expiration periods a configurable parameter.
  • Reduce the amount of memory used when a backup is terminated.

Ops Manager Server 3.4.8

Released on 2017-09-07

  • Allow session timeouts to be configured to a fraction of an hour.

  • Fix: Downloading the global diagnostic archive could fail if more than 32 MB of agent logs were stored.

  • Fix: Inconsistent Backup Configuration alert runs for config servers.

    Special Advisory

    After upgrading to Ops Manager 3.4.8, the Inconsistent Backup Configuration alert may be triggered for config servers. The backup for these config servers should be resynchonized.

    1. Click Backup.
    2. From the config server row, click on the Option menu [ ellipsis h icon ].
    3. Click Resync.

    Do not resync the entire sharded cluster if a config server triggers this alert. Resync the config server in question only.

  • Fix: Changes to the storage engine for a config server backup are applied on resync.

  • Fix: Validation for Backup Agent authentication mechanism failed if MONGODB-X509 was implicitly enabled.

  • Fix: In Local Mode, checks during Ops Manager startup allow configurations of deleted Groups to require MongoDB versions which are not present.

  • Security enhancements.

Ops Manager Server 3.4.7

Released on 2017-08-03

  • Agent Upgrades: Automation Agent 3.2.14.2187, Backup Agent 5.0.10.634.
  • Update JDK to 8u144.
  • Additional logging for oplog extensions due to delayed backups.
  • Additional logging for backup snapshots that retry due to unexpected file changes.
  • Fix: Ability to restore SCCC sharded clusters without a healthy mongos.
  • Fix: Race condition between snapshots and schedule update requests.
  • Fix: Do not query arbiters when executing an administrative error retrieval task.
  • Security improvements.

Ops Manager Server 3.4.6

Released on 2017-07-06

  • Agent Upgrades: Automation Agent 3.2.13.2141, Backup Agent 5.0.8.601.
  • Fix: Server Pool servers can get stuck in cleaning state if server is reimaged.
  • Fix: Make Backup ingestion collection stats calls idempotent.
  • Fix: Avoid skipping storage measurement rollups if customers have a database named hm.
  • Fix: Cause Discovery to update shard ID when shard is repurposed to arbiter.
  • Fix: Following upgrade from 2.x to 3.4.x, show progress of monitoring data migration.
  • Fix: Make all metrics available to cluster view.
  • Fix: Delay in Metric alerts check when opening and closing alerts for oplog metrics.
  • Fix: Metric alerts check may miss system measurements.

Ops Manager Server 3.4.5

Released on 2017-05-18

  • Agent Upgrades: Automation Agent 3.2.12.2107, Monitoring Agent 5.4.5.370, Backup Agent 5.0.7.494.
  • Ability to disable daily/weekly/monthly snapshots through the API.
  • Fix validation error for automated restores where the source or target cluster uses the storage.wiredTiger.directoryForIndexes parameter.
  • Fix validation errors when using LDAP Authorization along with LDAP and SCRAM-SHA1 authentication.
  • Various security enhancements.

Ops Manager Server 3.4.4

Released on 2017-03-30

  • Agent Upgrades: Automation Agent 3.2.11.2025, Backup Agent 5.0.6.486
  • Fix for premature forceful shutdown of the HEAD mongod instances for Ops Manager Backup Daemons running on Windows.
  • Fix problem with shutdown of mongod processes during an automated restores on Windows.
  • Fix for issues using Automation in multi-server deployments using SSL and encrypted PEM key files.
  • Fix for restores of backup snapshots stored unencrypted using file system storage.
  • Expose newest AWS regions as a storage option for S3 blockstores.
  • Optimization: Performance optimization for sending monitoring data to New Relic.
  • Security enhancement: Set X-Frame-Options: DENY HTTP headers for all authenticated requests.
  • Security enhancement: New option to allow enablement of HTTP Strict Transport Security (HSTS) with configurable maximum age.
  • Security enhancement: New option to allow use of browser session cookies, rather than time-limited persistent cookies.
  • Security enhancement: New option to disable all browser caching.

Ops Manager Server 3.4.3

Released on 2017-02-17

  • Agent Upgrades: Automation Agent 3.2.10.1997, Backup Agent 5.0.6.477, Monitoring Agent 5.4.4.366
  • Fix resource leak preventing successful backups when Ops Manager Backup is running on Windows.
  • Fix bug in removal of shards for sharded clusters on MongoDB 3.4.
  • Various security enhancements.
  • Support for running Automation, Backup and Monitoring Agents on MacOS Sierra.
  • Support for MongoDB Enterprise builds for Debian 7.1.
  • Fix import into Automation for deployments using LDAP authorization and SCRAM-SHA1 authentication.
  • Automation will deploy SSL-capable Windows builds for MongoDB 3.2.12+ and 3.4.2+.
  • Fix possible stall in the oplog application phase of Ops Manager Backups of MongoDB 3.4.

Ops Manager Server 3.4.2

Released on 2017-01-19

  • Agent Upgrades: Automation Agent 3.2.9.1985, Backup Agent 5.0.5.472
  • Fixed PageDuty notifications failing to send for System Alerts.
  • Fixed issue creating deployments with overlapping replica set / shard names of other deployments.
  • Fixed Global Alert UI validation being incorrect for some roles.
  • Fixed failure configuring MONGODB-CR + LDAP auth mechanisms + LDAP authz.
  • Added option to disable checksums during restores.
  • Upgrade to JDK8u121.

Ops Manager Server 3.4.1

Released on 2016-12-27

  • Agent Upgrades: Automation Agent 3.2.8.1942, Monitoring Agent 5.4.3.361, Backup Agent 5.0.4.469
  • Support for Ops Manager using Active Directory to authenticate to application databases.
  • Fixed roles info being lost when editting a database user.
  • Fixed issue starting backup when ‘net.compression’ is in use.
  • Prevent Ops Manager URLs from having a double forward slash when the URL configured by the user ends in a forward slash.
  • Fixed editing LDAP group mappings.
  • Fixed new OM 3.4.0 being unable to start for the first time without internet connectivity

Ops Manager Server 3.4.0

Released on 2016-11-29

General

  • Added support to monitor, back up, and automate MongoDB 3.4 deployments.
  • Added support for deploying Ops Manager on SUSE12.
  • Added support for Automation Agents, Backup Agents and Monitoring Agents on Ubuntu 16.x Power (ppc64le) and RHEL 7 Power (ppc64le).
  • Added support for mutual TLS/SSL between all Agents and Ops Manager.
  • Added the ability to assign tags to Projects and to filter global alerts by tags.
  • Added the ability to change a the name of a Project.

Automation

  • Includes more flexible handling of MongoDB User and Roles; specifically:
    • Ability to choose whether or not specific users and roles are managed, and
    • Ability to choose whether or not to allow MongoDB Users and Roles to be managed externally from Ops Manager.
  • Added the ability to manage sharded collections and zones.
  • Added support for MongoDB 3.4 LDAP Authorization.
  • Added support for Kerberos Authentication on Windows.

Backup

  • Added support for uncompressed WiredTiger snapshots on the filesystem.
  • Added support for storing snapshots in S3.
  • Added support for WiredTiger encryption at rest.
  • Added the ability to control the reference time for the snapshot schedule; e.g., specify that snapshots are taken every 6 hours, starting at 12:00:00 AM.
  • Added support for all data-format affecting MongoDB configuration options: directoryPerDB, smallfiles, etc.

Monitoring

  • Provides high resolution monitoring: metrics are now captured every 10 seconds.
  • Supports flexible retention monitoring: metrics at each resolution can be retained for a configurable amount of time.
  • Allows hardware metrics to be collected for any managed MongoDB process (i.e. any process that is managed by an Automation Agent).
  • Provides new API endpoint to retrieve all metrics for a particular MongoDB process in one API call.

Provisioning

  • Added support for maintaining a pool of provisioned servers from which users in a group can request servers to host new MongoDB deployments.

Release Advisories for Ops Manager 3.4

Before upgrading Ops Manager from 2.0.x to 3.4, complete the following actions:

  • Upgrade backing databases to at least MongoDB 3.0.8. MongoDB 2.6 is no longer allowed as a backing store.

  • Ensure the data partitions for the Ops Manager Application Database have at least 50% free disk space.

    Once the Ops Manager upgrade to 3.4 completes, Ops Manager begins migrating of all monitoring data to a new schema in the background. This migration requires significant free disk space.

  • Add necessary IP addresses or CIDR blocks for to the whitelists for any API clients connecting to the Automation configuration endpoints.

  • Back up the mms.conf file in your current installation if you have modified it.

    This is not a regular practice as mms.conf contains platform and network administration settings for Java VM and Ops Manager network port settings. You must use the new mms.conf file the upgrade installs. Reapply any modifications from your backed up copy to the new mms.conf after completing the upgrade.

For procedures to upgrade to 3.4, see Upgrade Ops Manager.

Ops Manager Server 2.0

Ops Manager Server 2.0.10

Released on 2017-08-03

  • Update JDK to 8u144.
  • Fix: Backup should compute size of incomplete snapshots.

Ops Manager Server 2.0.9

Released on 2017-05-23

Ops Manager Server 2.0.8

Released on 2017-01-19

  • Fixed minimum oplog check when starting backup from failing on RECOVERING members.
  • Fixed restores on expired, but not yet deleted, snapshots.
  • Upgrade to JDK8u121.

Ops Manager Server 2.0.7

Released on 2016-11-03

  • Agent Upgrades: Automation Agent 2.5.22.1876
  • Improve filesystem backup performance by increasing disk buffer size.
  • Fixed backup initial sync failures due to inefficient query on oplogs.
  • Fixed starting backup when MongoDB was using Kerberos auth and Automation was not in use.
  • Fixed old application settings appearing as overrides on the Ops Manager settings page
  • Fixed publish failure due to weakCertificateValidation improperly being set based on the allowConnectionsWithoutCertificates setting.
  • Support for Windows Server 2016 and Windows 10.
  • Upgrade to JDK8u112. Unlimited strength encryption policy included by default.

Ops Manager Server 2.0.6

Released on 2016-08-18

  • Agent Upgrades: Automation Agent 2.5.20.1755
  • Fixed case where acknowledged alerts could be opened again.
  • Fixed issue where DNS failures on the hostname(s) of the Ops Manager application database cause Ops Manager to shutdown.
  • File system snapshot stores can be used in group specific snapshot store filters.
  • Fixed issue where an unconfigured Backup Daemon could be assigned a backup job.
  • Upgrade to JDK8u102.

Ops Manager Server 2.0.5

Released on 2016-07-14

  • Agent Upgrades: Automation Agent 2.5.19.1732, Monitoring Agent 3.9.1.326
  • Fixed credentialstool on Windows, which is used to encrypt passwords in the config file.
  • Fixed Backup Daemon auto-download of RHEL platform specific builds.
  • Added support for LDAP referrals for Ops Manager user authentication.
  • Added support for changing LDAP search attribute for Ops Manager user authentication.
  • Fixed index creation UI in Firefox and IE11.

Ops Manager Server 2.0.4

Released on 2016-05-20

  • Agent Upgrades: Automation Agent 2.5.18.1647, Backup Agent 3.9.1.382
  • Fixed failure to generate diagnostics archive due to large amount of log data.
  • Validate Automation sslMode changes at publish time instead of draft.
  • Allow Automation to transition from sslMode disabled to not having a sslMode.
  • Fixed false positive auth mechanism validation failures when starting backup.
  • Fixed issue with processing some types of aggregation queries when calculating suggested indexes.
  • Fixed exception during a backup restore if that data previously was on a blockstore that has since been deleted.
  • Removed Ubuntu 14.04 enterprise builds for MongoDB 2.4.X that were erroneously in the version manifest.
  • Ability to edit LDAP Projects for Automation Admin was accidentally hidden.
  • Fixed javascript error on empty profiler view.
  • Upgrade to JDK8u92.

Ops Manager Server 2.0.3

Released on 2016-03-24

  • Agent Upgrades: Automation Agent 2.5.17.1604
  • Fixed critical bug in conversion to config server replica sets. Conversions to config server replica sets should be not performed with Ops Manager 2.0.2.
  • Fixed Ops Manager not recording HTTP access logs.
  • Fixed LDAP PEM settings from failing pre-flight checks even when LDAP wasn’t in use.
  • Fixed Automated Point-In-Time restores of Sharded Cluster with config server replica sets.
  • Fixed removing SSL for a Deployment
  • Email configuration changes no longer require a restart of the Ops Manager service.
  • Allow specifying a temporary port for use during conversion to config server replica sets.
  • Added Automation support for net.ssl.disabledProtocols.
  • Allow control over the compression level of the File System Snapshot Store.

Ops Manager Server 2.0.2

Released on 2016-03-01

  • Agent Upgrades: Automation Agent 2.5.16.1552
  • Added support rolling upgrades to config servers as a replica set (requires MongoDB 3.2.4+).
  • Added support for running Agents, the Ops Manager server, and MongoDB on SUSE12.
  • Added support for Slack and Flowdock notifications as system alerts.
  • Fixed Automation Admin Role missing from group LDAP configuration.
  • Fixed charting problem on Chrome 48+.
  • Fixed issue deleting processes that were part of a config server replica set.
  • Fixed issue where deployment drafts could prevent Ops Manager from starting in local mode.
  • Fixed issue where disabling 2FA in Ops Manager still required 2FA for users that had it configured.
  • Ops Manager upgrades on Windows require an uninstall of the previous version. This restriction was added to prevent issues that could occur on upgrade that are still unresolved without uninstall.
  • Upgrade to JDK8u74.

Ops Manager Server 2.0.1

Released 2016-01-21

  • Agent Upgrades: Automation Agent 2.5.15.1526.
  • Stability and performance improvements for restores via automation.
  • Support restores via automation for shared clusters with config server replica sets.
  • Fixed editing of managed users not promoting to re-enter password (Relevant only to imported SCRAM-SHA1 users.)
  • Fixed old errors from imports to automation impacting new imports.
  • Automation now updates the location of the keyfile according to the defined downloadBase.
  • Fixed cases where suggested indexes did not handle unexpected profiling data.
  • Fixed issue with filesystem snapshots failing when trying to resume a snapshot after restart.
  • Fixed LDAP form validation not allowing “ldaps”.
  • Fixed the Backup Daemon not recognizing Windows MongoDB Enterprise builds.
  • Fixed cases where global diagnostic archive would fail if it was too large.
  • Fixed importing into automation with SSL always requiring client certificates.

Ops Manager Server 2.0.0

Released 2015-12-08

General

  • Added support to monitor, back up, and automate MongoDB 3.2 deployments.

  • Single Ops Manager Package: there is no longer a separate package for the Backup Daemon. The single Ops Manager package installs both the Ops Manager Application and the Backup Daemon. You can configure any server with Ops Manager to handle backups through the Backup Admin interface.

  • Configuration in the Database: the Ops Manager application configuration is now stored in the application database rather than in configuration files. This allows for central configuration management.

    Each Ops Manager instance, on each server, must be configured with information on how to connect to the Application Database. Local config files override the information in the database: as such, switching to configuration in the database is not required, but is recommended.

  • Backup agent port change: Ops Manager no longer requires a separate port for backup traffic. All HTTP traffic is now over a single port. By default, Ops Manager uses port 8080.

  • Ops Manager 2.0 updates the admin interface to show the topology of Ops Manager software, the application database, and any backup databases.

  • Added support to convert to LDAP authentication for Ops Manager users at any time, with no downtime.

  • Upgraded to JDK8u66.

Automation

  • Added support for X.509 member authentication.
  • Improved handling of adding members to replica sets: to avoid disrupting majority writes, new members are now added to replica sets as priority=0, votes=0 until they reach secondary state, after which Ops Manager automatically updates the configuration to match the priority and votes value specified in the deployment.
  • Added the ability to manage indexes from the Ops Manager UI.
  • Improved index creation: indexes are now created in a rolling fashion.
  • Added Automation support for Windows MongoDB instances.

Monitoring

  • Added a new profiler with Suggested Indexes.
  • Added support for maintenance windows during which time Ops Manager does not send alert notifications.

Backup

  • Filesystem snapshot storage: added the ability to store snapshots on a plain shared file system instead of a MongoDB instance. With filesystem storage, Ops Manager stores snapshots in a directory hierarchy and the data files themselves are compressed using gzip.

  • Backup agent port change: Ops Manager no longer requires a separate port for backup traffic. All HTTP traffic is now over a single port. By default, Ops Manager uses port 8080

    Ops Manager will automatically update any Backup Agents managed by Automation to use the new port. You will need to manually update any Backup Agents set up manually after upgrading Ops Manager. The upgrade instructions describe how to configure the mothership field in the configuration files of non-automated Backup Agents.

  • Sync store no longer required: a dedicated sync store is no longer required: Backup initial syncs are “streamed” to the Backup Daemon and only use a small amount of temporary space in the oplog store.

  • Automated restores: added a new option to automatically restore a backup to a running replica set or sharded cluster.

  • Added support for namespace whitelisting, which allows you to back up only a subset of your data.

  • Added the ability to manage HTTP restore link expiration from the Ops Manager UI and through the API for each individual restore request.

  • Added support for the Backup Daemon to download required MongoDB binaries from the Ops Manager web server when they are not available locally.

Release Advisories for Ops Manager 2.0

Before upgrading to Ops Manager 2.0, ensure the following are true:

Backup Database

Databases that store the Snapshot Storage must be online.

There are data migrations that touch the various backup data stores that make up the Snapshot Storage. Any data stores that are no longer in use should be deleted through the Ops Manager UI before upgrading.

Backup Daemon

The Ops Manager package installs the Backup Daemon. When started, the Ops Manager package automatically starts two services: the Ops Manager Application and the Backup Daemon. You choose on which hosts to “activate” the Backup Daemon. The daemon always runs, but performs no operations until activated.

After upgrading to 2.0, a host that runs only the Ops Manager Application continues also runs a “dormant” Backup Daemon service. The Backup Daemon remains dormant as long as you do not activate it.

A host that runs only a Backup Daemon runs Ops Manager with an “activated” Backup Daemon and a “dormant” Ops Manager Application. The Ops Manager Application remains dormant as long as you do not direct HTTP traffic to it.

Backup HTTP Service

Change the mothership setting in all Backup Agent configuration files for any manually installed agents to use port 8080. Any manually installed Backup Agents that are not set to use port 8080 cannot connect to Ops Manager You must have access to the hosts running any manually installed Backup Agents.

Beginning with Ops Manager 2.0, there is no Backup HTTP Service on port 8081. Any Backup Agents that Automation manages are updated to use the new port (8080) automatically.

Agent Updates
The agents have not been updated before the Ops Manager Application has been upgraded. If you use Automation, Ops Manager prompts you to update the agents after you upgrade. Follow the prompts to update the agents through the Ops Manager Application. Do not update the agents manually.
conf-mms.properties
Beginning in 2.0, Ops Manager stores global configuration settings in the Ops Manager Application Database and stores only local settings in the Ops Manager host’s conf-mms.properties file. The upgrade procedure uses the existing conf-mms.properties file to connect to the Ops Manager Application Database before replacing the existing file with the new, smaller 2.0 file.
Restore properties

The following properties no longer apply and are replaced by settings specified when initiating a restore:

  • mms.backup.restore.linkExpirationHours
  • mms.backup.restore.linkUnlimitedUses
  • If you have modified the mms.conf file in your current installation, you have backed it up.

    This is not a regular practice as mms.conf contains platform and network administration settings for Java VM and Ops Manager network port settings. You must use the new mms.conf file the upgrade installs. Reapply any modifications from your backed up copy to the new mms.conf after completing the upgrade.

See: Upgrade Ops Manager for upgrade instructions for your operating system.

Ops Manager Server 1.8

Ops Manager Server 1.8.3

Released 2015-12-15

  • Fixed issue where monitoring settings for existing servers were not always editable.
  • Support for additional Amazon Simple Email Server regions. To specify regions other than the default US-EAST, see aws.ses.endpoint.
  • Fixed SNMP notification mechanism for System Alerts.
  • Fixed user privileges for MongoDB 3.0 missing from UI on Users & Roles page.
  • Upgraded to JDK8u66.

Ops Manager Server 1.8.2

Released 2015-10-20

Ops Manager Server 1.8.1

Released 2015-08-17

  • Agent Updates: Automation Agent 2.0.12.1238, Monitoring Agent 3.7.0.212
  • Updated Backup seedSecondary script for MongoDB 3.0.
  • Fixed adding users with GLOBAL roles to individual groups.
  • Fixed Host Down alerts not firing correctly for arbiters.
  • Fixed error when trying to enable X.509 authentication for Monitoring only (without Automation).
  • Fixed error when trying to enable host log collection.
  • Fixed case where an acknowledged Alert can be re-opened when Alert processing is behind.
  • Fixed case where monitoring classified a Config Server as a Standalone when there were no mongos services.

Ops Manager Server 1.8.0

Released 2015-06-23

Security

  • Automation now supports SSL and MongoDB Enterprise authentication mechanisms: Kerberos, LDAP, and X.509.

    Ops Manager 1.8 can start new MongoDB instances using SSL and enterprise authentication mechanisms and import existing instances using SSL and enterprise authentication for management.

  • Added the ability to specify a proxy server for Ops Manager to use to access external services.

  • Added support for self-signed CAs and client certificates when using SSL LDAP for Ops Manager user authentication.

Alerts

  • System Alerts: system alerts allow an Ops Manager Administrator to receive alerts when the state of the software itself is unhealthy.
  • Global Alerts: global alerts allow an Ops Manager administrator to monitor any set of Ops Manager groups without needing to configure the alerts on a group-by-group basis.
  • Added the ability to deliver Project, Global, and System alerts via an HTTP webhook.
  • Lowered the alerting check frequency from five minutes to one minute, allowing for more responsive alerts.

Automation

  • Automation now uses distribution-specific builds for MongoDB Community Edition when one is available for the operating system and version in use. Previously, Automation used the generic MongoDB Community Edition build.

    Upgrading the Automation Agent and Ops Manager to the new version will not automatically change your MongoDB deployments to a distribution- specific build: if you wish to use the distribution-specific build, you will need to update the MongoDB version.

  • Added support to change the storage engine for a MongoDB deployment using Automation.

  • Beta: Added Automation support for Windows MongoDB instances. This feature must be enabled for an Ops Manager group for it to be available.

Monitoring

  • Standby Monitoring Agents now check in with Ops Manager more frequently. You can now configure the Monitoring Agent session timeout to allow for faster failover. See: Monitoring Agent Session Failover for more information.

Backup

  • Added the ability to configure the Backup Database’s block size. The Configure Block Size in a Blockstore tutorial describes how to configure the size of the blocks in the Backup Database’s blockstore.
  • Added the ability to initiate backup SCP restores through the Public API. See: Restore Jobs.

Considerations for Upgrade (v1.8)

  • Ops Manager 1.8 requires that the Ops Manager Application Database and Snapshot Storage run MongoDB 2.6 or later. Ops Manager will not start after upgrade if your backing databases are using an earlier version of MongoDB. The MongoDB Manual provides upgrade tutorials with each release. To upgrade from MongoDB 2.4 to 2.6, see: Upgrade MongoDB to 2.6.

  • When you upgrade to Ops Manager 1.8, Ops Manager disables all Automation Agents until they are upgraded to Automation Agent 2.0.9.1201. You can upgrade the Automation Agents by clicking the link that appears in the Please upgrade your agents banner that will appear on the Deployment page in the Ops Manager interface.

  • Direct upgrade is only allowed from Ops Manager 1.5 and Ops Manager 1.6. To upgrade to Ops Manager 1.8 from an earlier version of MongoDB, you must first upgrade to Ops Manager 1.6, and then to 1.8.

  • In Ops Manager 1.8, mms.multiFactorAuth.level replaces the deprecated mms.multiFactorAuth.require setting. mms.multiFactorAuth.level supports more values than its predecessor.

    Ops Manager will not start with mms.multiFactorAuth.require in the properties file, but will report an error indicating that the setting has been deprecated, and that you must update your configuration.

  • Ops Manager 1.8 does not include the Backup HTTP Service: its functionality is now part of System Alerts and Global Alerts.

  • System Alerts give new insight into the health of Ops Manager and may immediately trigger on upgrade if Ops Manager is not in the expected state. For example, if your Application or Backup databases have startup warnings or if the connection strings to those databases point to any unreachable MongoDB instances, Ops Manager will issue an alert.

  • The Ops Manager Deployment user interface has been streamlined such that the View Mode and Edit Mode dual views have been merged into a unified view.

Ops Manager Server 1.6

Ops Manager Server 1.6.4

Released 2015-08-17

  • Ops Manager no longer shuts down if the Ops Manager Application Database is unreachable. (This issue was erroneously reported as resolved in Ops Manager 1.6.3.)

Ops Manager Server 1.6.3

Released 2015-06-23

  • Agent updates: Automation Agent 1.4.18.1199-1
  • Added full support for restores of WiredTiger backups. Previously, Ops Manager only supported SCP Individual File restores for WiredTiger backups.
  • Added optimization to prevent some Backup Daemon background tasks from doing excessive logging when databases are down.
  • Fixed a user interface issue when displaying an empty Automation diff.

Ops Manager Server 1.6.2

Released 2015-04-28

  • Fixed issue with grooms on a WiredTiger Backup Blockstore.
  • Fixed a possible connection leak with the SCP Individual File restore type.
  • LDAP users are now periodically synced with the LDAP server to prevent communications after a user is removed from a group.
  • Fixed an issue with backups of MongoDB 3.0 mongod instances running with the --setParameter failIndexKeyTooLong=0 option.

Ops Manager Server 1.6.1

Released 2015-03-26

  • Agent updates: Automation Agent 1.4.15.999.
  • Security Update: resolved an issue where users removed from LDAP groups were not always removed from corresponding Ops Manager groups. This upgrade is highly recommended for anyone using LDAP authentication.
  • Selecting wildcards in the Version Manager is no longer supported when automation.versions.source is set to local.
  • Added a 1 hour timeout to kill a Backup head database if it does not shutdown cleanly. You must perform a resync following a hard kill.
  • Windows support for Backup Daemon using Windows 64-bit 2008 R2+ MongoDB builds.
  • Fix for Backups stored in WiredTiger format in which a single collection grows from under 8 GB to over 8 GB in size.
  • The time before an unreachable mongos process is deactivated is now configurable on a per group basis. See Admin-Only Group Settings.
  • The time before a standby Monitoring Agent takes over after the primary Monitoring Agent stops responding is now configurable to a minimum of 90 seconds. See the mms.monitoring.agent.session.timeoutMillis setting in Ops Manager Configuration.
  • For Backup HTTP pull restore, the link expiration and the number of allowed uses of a link are now configurable.

Ops Manager Server 1.6.0

Released 2015-03-02

New Features

  • Initial release of Automation. Automation manages many basic administrative tasks for MongoDB deployments, including version upgrades, adding replica set members, adding shards, and changing oplog size. You can both import existing deployments into Automation and create new deployments on your provisioned hardware.

  • Windows support (Monitoring and Backup only). You can Install Ops Manager on Microsoft Windows using MSI files. Ops Manager supports Windows Server 2008 R2 and above.

  • Support for MongoDB 3.0, including support for backups that use the WiredTiger storage engine.

    To monitor or back up MongoDB 3.0 deployments, you must install Ops Manager 1.6 or higher. To monitor a MongoDB 3.0 deployment, you must also run Monitoring Agent version 2.7.0 or higher.

  • Support for using the SSL and MONGODB-X509 authentication mechanisms for the backing MongoDB databases. See Configure the Connections to the Backing MongoDB Instances.

  • Public API endpoints to manage Automation configuration. For more information, see Automation in the API documentation.

Improvements

  • The Ops Manager’s Administration interface provides more information to make it easier to monitor the health of the Ops Manager installation.
  • The Ops Manager Deployment tab now displays all deployment information on one page, with icons for selecting view options. The new Topology View groups all hosts by the replica set or sharded cluster they are part of. The new Servers View shows information about MongoDB processes and Ops Manager agents grouped by server.
  • Fixed an issue (MMS-2273) where, in certain situations, the Backup Agent was not reporting a cluster snapshot as potentially inconsistent.
  • Improved handling of cursor timeouts by the Backup Agent. To use this improvement, upgrade to the latest Backup Agent, which is included with Ops Manager. The improvement became available with Backup Agent 2.3.3.209-1.

Considerations for Upgrade to 1.6

  • Ops Manager 1.8.0, when released, will not support MongoDB 2.4 for the Ops Manager Application Database and Snapshot Storage. Ops Manager Server 1.8.0 will continue to support MongoDB 2.4 for your monitored and backed-up databases.
  • Ops Manager 1.6.0 supports direct upgrades only from MMS On Prem 1.3 and above.
  • The procedure to configure Ops Manager to run with HTTPS has changed and is greatly simplified. The previous procedure no longer works. For the new procedure, see Configure SSL Connections to Ops Manager.
  • The connection string to the Backup Blockstore database is now configured through the Administration interface’s Blockstores page and not through the mongo.backupdb.mongoUri field in the conf-daemon.properties configuration file.
  • Ops Manager no longer requires you to supply the replica set name of the backing MongoDB instances. The mongo.replicaSet and mongo.backupdb.replicaSet properties have been removed from the configuration files. These properties had previously controlled whether Ops Manager treated a connection to a backing instance as a standalone or replica set, for the purpose of setting the write concern. Ops Manager now sets write concern based on how many hosts are supplied in the connection string.
  • You can disable Automation for the entire Ops Manager installation through the mms.featureFlag.automation setting in the conf-daemon.properties configuration file.
  • Removed the Dashboards view from the Ops Manager UI. You can view monitoring metrics from the Deployment tab. See: View Metrics for an overview of the available metrics and how to access them.

MMS Onprem Server 1.5

MMS Onprem Server 1.5.5

Released 2015-03-26

  • Security Update: resolved issue where users removed from LDAP groups were not always removed from corresponding Ops Manager groups. This upgrade is highly recommended for anyone using LDAP authentication.

MMS Onprem Server 1.5.4

Released 2015-03-18

  • Fixed race condition that could cause the Backup Daemon to hang when the MongoDB process for a head database fails to start.
  • Fixed an issue where a rollback occurring shortly after a terminate could step on the terminate.
  • The time before an unreachable mongos process is deactivated is now configurable on a per group basis. See Admin-Only Group Settings.
  • The time before a standby Monitoring Agent takes over after the primary Monitoring Agent stops responding is now configurable to a minimum of 90 seconds. See the mms.monitoring.agent.session.timeoutMillis setting in Ops Manager Configuration.
  • For Backup HTTP pull restore, the link expiration and the number of allowed uses of a link are now configurable.

MMS OnPrem Server 1.5.3

Released 2014-12-17

Significant improvements in performance for the processing of MMS OnPrem Monitoring data for MMS OnPrem Projects with a large number of hosts

MMS OnPrem Server 1.5.2

Released 2014-11-18

  • Added Support for archive restores (.tar.gz) for databases whose filenames exceed 100 characters.
  • API: Skip missed points in metrics data, instead of returning empty data.
  • API: Return correct number of data points when querying metric data with the period option.
  • Backup Agent update to 2.3.3.209-1

MMS OnPrem Server 1.5.1

Released 2014-09-26

  • Fix cases where replica set member alerts (e.g. no primary, number of healthy members) could send false positives.
  • Skip backup-daemon rootDirectory and mongo.backupdb.mongoUri overlap check when the mongo.backupdb.mongoUri is on a different host.
  • mms-gen-key script handles user’s effective group being different than the username.
  • Security enhancements.

MMS OnPrem Server 1.5.0

Released 2014-09-02

Considerations for Upgrade

  • MMS OnPrem only supports direct upgrades from 1.3 and 1.4.

  • Change in configurations and policy for 2FA: Two-factor authentication must now be explicitly enabled using the mms.multiFactorAuth.require setting.

  • The default LDAP group separator became ;;. Previously the separator was ,. See the LDAP configuration documentation for more information.

  • Suppressed hosts will only remain suppressed for 30 minutes.

    Previously, if after deleting a host, from MMS OnPrem Monitoring the hostname and port combination would be added to a suppression list with an infinite lifetime. The suppression list prevented a race condition where host in a cluster would be auto-discovered by another member of a deployment before the host could was fully removed. Now, hostname and port combinations remain on the suppression list for only 30 minutes.

  • Set the mms.remoteIp.header in the conf-mms.properties file if clients access the MMS OnPrem Application via a load balancer.

  • mongo.backupdb.mongoUri is no longer in conf-mms.properties. This was previously a required field in this file. It remains in the backup daemons’s conf-daemon.properties.

  • Stored MongoDB profile data is not transferred between OnPrem 1.4 and OnPrem 1.5 during the upgrade process.

Improvements

  • When an MMS OnPrem Backup job fails to bind, the system periodically and automatically retries.
  • All MMS OnPrem Backup jobs will retry indefinitely.
  • Point in Time restores are now available with one second granularity.

New Features

MMS OnPrem Server 1.4

MMS OnPrem Server 1.4.3

Released 2014-07-22

  • Addressed issues related to Backup Job assignment for 2.6.x clusters that used the clusterMonitor role to support MMS OnPrem Monitoring.
  • Fixed problem importing email addresses for users for deployments that use LDAP integration.
  • Fixed rare race condition caused high CPU usage in the MMS OnPrem HTTP Service if the application cannot connect to one of the backing databases.
  • Additional security enhancements.

MMS OnPrem Server 1.4.2

Released 2014-05-29

  • Critical bug fix for backing up MongoDB 2.6 deployments that include user or custom role definitions:

    • The system.version collection in the admin database will be included in all future snapshots.
    • The system.roles collection in the admin database will be included after a new initial sync is performed.

    Users capturing backups of MongoDB 2.6 replica sets or clusters with MMS OnPrem that include custom role definitions should perform a new initial sync. Taking a new initial sync will ensure that the role definitions are included in the backup.

  • Disable MongoDB usePowerOf2Sizes for insert-only MMS OnPrem Backup collections.

  • Speed optimization for MMS OnPrem Backup HTTP pull restores.

  • Fix for LDAP integration, MMS OnPrem now passes full dn correctly when authenticating the user.

MMS OnPrem Server 1.4.1

Released 2014-04-28

  • Ability to Backup replica sets or clusters using Kerberos authentication.
  • Ability to Backup replica sets or clusters running specific custom MongoDB builds provided by MongoDB, Inc.
  • Fix validation issue preventing Backup of MongoDB 2.6.0 clusters.
  • Reduced log noise from Monitoring Agent when monitoring MongoDB 2.0 or unreachable mongods.

MMS OnPrem Server 1.4.0

Released 2014-04-08

  • Includes MMS OnPrem Backup: continuous backup with point-in-time recovery of replica sets and cluster-wide snapshots of sharded clusters.
  • Finer-grained roles and permissions.
  • Improved user interface for alerts.
  • Enhanced Activity Feed for auditing of all activity.
  • Monitoring Agent distributed as OS-specific binary. Python dependency removed.
  • LDAP integration for managing users and groups.

MMS OnPrem 1.4.0 requires MongoDB 2.4.9+ instances for backing storage.

MMS OnPrem Server 1.3

Released 2013-12-01

  • Packaging/support for Debian and SUSE Linux.
  • Kerberos authentication support between MMS OnPrem server and backing MongoDBs, as well as between Monitoring Agent and the MongoDBs it monitors.
  • OnPrem users can be overall site administrators. (MMS OnPrem Admins)
  • New admin section where MMS OnPrem Admins can manage user roles and message banners.
  • Tunable advanced password and session management configurations.
  • Encryption key rotation, more specific CORS policy, auth tokens removed from chart URLs, and other security enhancements.

MMS OnPrem Server 1.2

Released 2013-07-24

  • Redesigned user interface and enhanced algorithm to auto-discover hosts and derive host topology.
  • SNMP monitoring.
  • Ability to export charts.
  • Option to store encrypted authentication credentials in the mmsDb property in the configuration file.
  • Ability to classify users within an MMS OnPrem Project as group administrators or read-only users.