Internal Integration Release Notes

CAPI: Added new metrics and fixed vulnerabilities.

Application (Java) Self Monitoring: Vulnerability fixes.

Public API (PAPI): Vulnerability fix and a new API added.

Virtana Cloud Sync: Vulnerability fixes.

Dynamic Alarm: Vulnerability fix and bug fixes.

KVM: Vulnerability fixes.

Linux: Metrics improvements and vulnerability fixes.

Solaris: Vulnerability fixes.

UCS: Vulnerability fixes.

UCP: Vulnerability fixes.

Windows: Vulnerability fixes and metrics improvements.

Cisco Nexus: Entity and metrics improvements, plus vulnerability fixes.

IP Switch: Bug fix, metrics improvements, and vulnerability fixes.

Hitachi VSP: Vulnerability fix and bug fixes.

Hitachi NAS: Bug fix and vulnerability fix and new metrics added.

Hitachi Content Platform: Vulnerability fixes.

VPLEX: Vulnerability fixes.

Hitachi VSP: Vulnerability fix and bug fixes.

Hitachi NAS: Bug fix and vulnerability fixes.

Hitachi Content Platform (HCP): Vulnerability fixes.

VPLEX: Vulnerability fixes.

DataDomain: New release.

HPE 3PAR: Vulnerability fix and bug fixes.

Infinibox: Vulnerability fixes.

Isilon/PowerScale: Vulnerability fixes.

NetApp: Metrics added and vulnerability fixes.

PowerMax: Vulnerability fixes.

PowerStore: Vulnerability fixes.

Pure: Vulnerability fix and bug fixes.

Nutanix: Vulnerability fix and bug fixes.

SVC: Vulnerability fix and bug fixes.

Unity: Vulnerability fixes.

XtremIO: Vulnerability fixes.

Linux: Added new MTU metrics and addressed vulnerability issues.

Solaris: Fixed security vulnerabilities.

UCS : Released shell version fix.

Windows : Resolved vulnerability issues.

Cisco Nexus : Improved entity handling and enhanced metrics collection.

IP Switch: Bug fixes and metrics improvements.

Hitachi VSP: Bug fix implemented.

Hitachi NAS: Bug fix implemented.

Vplex: Shell update applied.

HPE 3Par: Shell update released.

Isilon/Powerscale: Security fixes.

NetApp: Added new metrics.

PowerMax: Security fixes.

Powerstore: Shell update applied.

Pure: Added new entities and metrics.

Nutanix: Bug fix implemented.

SVC: Security fixes and Branding updates.

Unity and Xtremlo: Shell update released.

CAPI, PAPI, VCS, Dynamic Alarm: Security fixes.

Cisco Nexus: Bug fixes and metrics improvements.

HCP: Added new properties.

IP Switch: Bug fixes and metrics improvements.

Isilon/PowerScale: Bug fixes.

Linux: Bug fixes and metrics improvements.

NAS: Bug fixes, metrics, and entity improvements.

NetApp: Performance improvements.

Nutanix: Bug fixes.

PAPI: New APIs added, bulk Create/Read/Update/Delete operations on Windows and Linux host subscribed in the OS integrations.

Powermax: Bug fixes.

Pure: Added new properties.

UCP: New integrations.

UCS: Bug fixes.

VCS: Bug fixes.

VSP: Pool performance metrics.

Windows: Bug fixes, metrics improvements, and performance improvements.

Hitachi NAS: Bug fixes and metrics improvements.

HPC: Added new properties.

InfiniBox: : Metrics improvements.

IP Switch: Bug fixes and metrics improvement. Added Nvidia cumulus switch support.

Isilon/PowerScale: Bug fixes.

KVM: Bug fixes.

NetApp: Bug fixes.

Linux : Bug fixes and metrics improvements.

Nexus: Bug fixes.

Nutanix: Bug fixes and CVM changes.

PowerMax: Java updates.  

Public API: Bug fixes.

Pure: Added new properties.

IBM Mod: iSCSI improvements and added new metrics.

UCS: Bug fixes.

VSP: Bug fixes.

Windows: Bug fixes, Windows 2022 support, metrics improvements, and PS7 support.

Shell Updates: Implemented updates for Hitachi VSP, VPLEX, HPE3PAR, Infinibox, PowerStore, Unity, and XtremIO.and IBM Mod.

Java Update: Updated Java components for Isilon, NetApp, PowerMax, and IBM Mod.

Bug Fixes: Addressed issues in Pure and Nutanix integrations.

The integration now supports obtaining the IQN (iSCSI Qualified Name) address from ESX Host discovery. The system automatically creates an internal entity called iSCSIAddress", which can be used for linking iSCSI storage arrays, such as NetApp and Pure.

Keycloack authentication: IO now supports authentication using Keycloak tokens provided by Global View. When a user logs in to the IO tab within Global View, the Keycloak token is validated with the Keycloak server, and IO creates its own session token for the user.There are two configurable UI service properties for customizing the Keycloak server URL and client ID:

com.vi.keycloak.server.url

com.vi.keycloak.client.id

By default, these are set to the VP Keycloak server settings, and modifications are only needed if you wish to connect to a different server.Additionally, IO users will be automatically created if they do not already exist. When a new user is created during Keycloak login, the default role assigned will be "io-readonly." This default role can be changed using the UI property com.vi.keycloak.newuser.role, with the options being "io-user," "IO-admin," or "io-readonly."

The action service now supports multiple target configurations and includes an HTTP action type.

Kubernetes Installation and Upgrade added for Kubernetes.

Linux : If this used on 7.2.6 appliance, new property will show, but it will not show topology for iscsi. If this integration is used on 8.0.0 appliance, it will show property and also topology related to iscsi.

Netflow integration cannot be correctly configured in kubernetes environment.

Nutanix:  

Storage Node provisioned metric value is shown as 0 since API does not provide any valid value.

Some of APIs are returning incorrect values causing UI to show incorrect values for capacity metrics of virtual disk and VM.

CAPI: Core Update and New Metrics

Application (Java) Self Monitoring: ROI Metrics added: New Metrics include

Public API (PAPI): New APIs introduced to support Grafana integration.

Virtana Cloud Sync: Removed dependency on application creation for sending alarms to the Virtana Platform (VP).

Linux: New properties added for hosts

Windows: New properties added for hosts

Hitachi VSP: New metrics and entities added.

Hitachi NAS and HCP: Beta release with new features.

HPE 3PAR and Unity: Bug fixes implemented.

PowerMax: New relationships added for improved monitoring and management.

Slack App: Action service integrated, allowing for automated actions via Slack.

Security Updates for: Infinibox, Isilon, NetApp, PowerMax, Power Store, Pure, Unity, Public API (PAPI), Kubernetes, Solaris, UCS,IBM Mod, HPE3PAR, XTREMlO, VPLEX

KVM: Data collection using Prometheus.

Linux: New metrics added.

Multipath version, HBA model, manufacturer, part no, product name, driver version

Windows: New properties added·  

Driver version, firm version manufacturer and model properties.

Cisco Nexus / IP Switch : New metrics added

Hitachi:  New metrics added.

Slack app: new command updates.

(Isilon): Bug Fixes.

(PowerMax): Bug Fixes.

(svc): Support and Licensing changes for IBM Flash Series 9000.

(Nexus): Bug fixes.

(Pure): New entity added: "Pure Flash Host Group" and its entity, New property added: "Software Version" for the "Pure Flash Array" entity. New property added: "Serial Number" for both "Pure Flash Array" and "Pure Flash Controller" entities.

(KVM) : Added OpenStack Support.

(Windows) and (Linux) : New metrics added- Ethernet Network Sent and Received Utilization (%) , Storage Avg Read vs Write Bytes Ratio, CPU Queue Length Real Memory (%) Cache Memory Committed Memory

(PAPI) : Bug fixes.

(IPSwitch): Bug fixes.

(Hitachi): New metrics added -

Dynamic Provisioning Efficiency %

Pool Physical Capacity Usage Rate

Pool Physical FMC Capacity Utilization %

Pool Subscription Rate

Pool Usage Rate

Pool Relocation Status

% Write Pending for Hitachi Microprocessor Blade

Max % Write Pending for Hitachi Microprocessor Blade

% Write Pending for Hitachi Storage System

Max % Write Pending for Hitachi Storage System

System CPU Time % for LdevMax System CPU Time % for Ldev

Max System CPU Time % for Hitachi Microprocessor Blade Core

Storage Max Read or Write Bytes for Port

System CPU Time % for Parity Group

(Infinibox) : Added as part of Uber bundle.

(Isilon) IO retrieves metrics in specific time intervals with a fixed number of data-points for Total Connected Clients. Isilon's API sometimes sends less number of data-points than expected for a requested time frame which creates a discrepancy on IO's part. To tackle this issue IO will use the last received data point's value and use them for the missing data-points.( usually we are just missing 1 data-point)

(PowerMax) In the newer versions of PowerMax/Vmax and with Unisphere 10, many changes have been done with respect to software and hardware by Dell. Below are some of the behavioral changes in the VMAX Integration Affected Models : V4 Arrays [POWERMAX 2500, POWERMAX 8500]

Affected Unisphere : 10+

Discovery Behaviour : The newer models specifically arrays with microcode 6079 have hardware changes. Hence the discovery on the affected models won't be collected for the following entities : [BEPort, CachePartition]

Metric Collection : No metric data for the following entities : [BEPort,CachePartition]

Below metrics are unsupported for the V4 arrays and 10 + unisphere :

BE Director : Busy Logical Core Metrics

FE Director: Busy Logical Core Metrics

Array Metrics : Write Pending Count, System Maximum Write Pending Limit

Note : The collection behaviour for Capacity Metrics :["OverallCompressionRatio","PercentSubscribedCapacity", "UsableCapacity", "SubscribedCapacity", "AllocatedCapacity", "PercentEffectiveUsedCapacity"] is refactored.

Also No Default Cache Partition in Powermax 2000 and Powermax 8000 models. So Cache not to be discovered.

(UCS) Bug fixes.

(Cisco Nexus) Support for telemetry and other bug fixes.

(Windows) Fixed issue with Error counter reporting

(VCS) Data collection performance improvements.

(VSP) Added CPU metrics to Hitachi Storage Systems, Hitachi Microprocessor Blades and Hitachi Microprocessor Blade Cores

(CAPI) Updates for Cloud CAPI / Virtana Cloud Sync.

(IP Switch) Fixed typo in the Container title to make it plural.

(Infinibox) Fixed missing IOPS metrics for some arrays. Also fixed the issue where Infinibox stops collecting data due to SQL error: (167, "Out of range value for column \'rowid\' at row 1")\

HPE (3PAR) Changed how Array Storage Throughput metrics are calculated. Array Storage Throughput metrics are aggregated based on port metrics.

(Isilon) Provided support for IO 7.2.2.

(NetApp) Fixed an issue which was preventing some metrics from aggregating correctly.

(PowerMax) Update for storage view. Handled capacity metrics for unisphere software v10.X. Also Fixed issue preventing storage metrics if some metrics were missing (due to unisphere version changes). Partial metric responses are now supported.

(Pure) Provided property "Skip Entity Type" to disable the creation of "Snapshots". This would avoid the discovery of 100,000 snapshots in short period.

(IBM Mod) Provided support for IO 7.2.2

(Cisco Nexus) Support for telemetry and other bug fixes

(CAPI) All integrations are updated to a new CAPI header format to handle a security update for CloudMonitor

All integrations require IO 7.2.1 as the base core version

(Linux) Fixed sporadic issue causing data collection to stop

(UCS) Updates to an internal library used to connect to the UCS Manager

(HVSP) Fixed a sporadic discovery issue due to REST API

(Infinibox) Fixed issue causing shutdown of translator after start

(Powerstore, Unity, XtremIO) Add global read timeout

(CAPI) Fixed issue causing some translator publishes to overwrite each other.

(CAPI) Added functionality to auto prune unused docker volumes when start/stopping translator containers.

(Self Monitor Java) Version bump.

(Nexus) Fixed Nexus9K data collection after restore of 4210 lab portal to a new distributed VM lab portal.

(PowerMax) Dynamically select 10.0, 9.2, 9.1 or 8.4 rest API when performing discovery based on supported Unisphere version.

(CAPI) Improved VCS registration time when IO connects to VP.

(Uber Bundle) There may still be sizing issues with this month's uber bundle.

(CAPI) Fixed a DIO issue.

(CAPI) log4j2.xml will not loose any custom log settings upon service restart.

(CAPI) Will now clean up containers from previous releases.

Log4j security-related fix: Security scan found a vulnerable unused jar that was introduced in September's releases. Fix was to remove unused jar from build manifests of the CAPI/OI integrations. No functional changes.

DIO fix: CAPI/Open Integrations failing to start on DIO 7.1.2+ requiring a manual work around. This has been fixed.

The UCS fix includes logic that is configurable via one or more properties. Reach out to support for more information.

Changes in topology and reporting for Storage Integrations. Displaying direct links from storage arrays to disks and disks groups (if they exist for a given integration).

The primary thrust of updates this month are packaging updates, which will reduce the size of the uber bundle by around 40%. Please note for that this release, there are still integrations that are packages separately from the uber bundle; see above.

Updating audit logs with the CRUD actions and additional details of Capi and translator configurations.

Fixed an issue causing the Entity Type property column to be named with an integration-specific type name, which also persisted through export.

Fixed several Snyk security-related issues.

The Public API version 1.2.0 is GA with this month's releases. It is available for 6.8.3+ releases. There is a new set of release notes for the API and, for this month only, the UG content, which won't be updated until the September 2022 release, is part of those notes.

Fixed Permission Denied issue with the show-heap-sizes Services and Support script. This change is in CAPI, IBM Mod , ISILON, PowerMax, and Self-Monitoring.

Security update: upgraded spark-core library (applies to all integration services)

Support for group combo, which the Nexus integration needed.

Pure: Enabled (in CAPI) migration to remove duplicate Pure entities. CAPI property to disable: com.vi.integrations.capi.enable_legacy_key_migration_375=false

A new, early access (beta) public API is available for 6.8.1 through 7.1.0 releases. Reach out to Product Management if you have a customer that might want to take a look. Documentation will be releases as part of the September 2022 Infrastructure Observability release.

Reduced logging generated by zeroing out metrics.

Added new self-monitoring metrics for Vertica: query counts and elapsed time.

Numerous CAPI bug fixes.

As noted above, several integrations are in a "controlled release" status, meaning Engineering assistance will likely be needed to deploy. Reach out to Product Management for coordination with Engineering.

Fixed an issue where integrations still collect data when the last instance is moved to the disabled state.

(PowerMax) Fixed an issue preventing the “Write Pending Count” sync replication metric from showing.

Self-Monitoring has been updated with two new graph store only metrics:

Both require 7.0.2+ core release.

CAPI 3.4.0 - a few internal bug fixes; clean up of duplicate entities.

A new SSL certificate fetch feature was added to core CAPI. Eventually, all integrations that require SSL certificates will share the same user experience, but for this release, it was added only to the PowerMax integration. A new shared chapter called SSL Certificate Handling contains the documentation for this new feature; it's part of the Creating a New Discovery Instance chapter, now in the PowerMax User Guide.

Licensing was updated. There are now two principal licensing mechanisms: per-configuration and per-device. A new License Validation chapter has been added to all of the translator-based integration user guides with details. Please also refer to the TOI content and recording.

Global change made to filter out erroneous latency values for storage throughput. Please note: this fix, while global in nature, was targeted to the KVM integration, whose March release notes have been updated to direct customers to the April release bundle.

Fixed a bug that caused the FibreChannelPath to Storage and FibreChannel Zoned Storage topology views go missing when latest Pure Integration was installed. This may have affected other storage-related integrations as well.

If the com.vi.integrations.capi.optimizedDiscoveryRequests.drop_structure_only_min property was set by Engineering or Support, it could impact a newly installed integration's ability to receive its first publish. To workaround this, temporarily set the property to 0 to allow the first publish to proceed. Validate that discovery was successful. If so, you can return the property to its previous value.

Fixed a sporadic issue with failed collections due to firewall after Appliance restart.

There is now a single install bundle that updates all integrations.

(CAPI internal) Normalize WWNs published through translator integrations. WWNs are now presented in a standard format and case to provide consistency with upstream.

Enhanced Infrastructure Observability topology views to support entity unification. Please note: integrations won't be able to take advantage of this feature until Infrastructure Observability release 6.8.3. There will be a separate TOI on this feature when 6.8.3 releases.

All affected integrations have been updated to address CVE-2021-44228, CVE-2021-45046, CVE-2021-45105, and CVE-2021-44832 by updating Apache Log4j to version 2.17.1.

The following items are forward-looking and do not affect this release:

Finally and without further commentary: Device level trace is disabled system wide. Please enable to Capture Full trace.

Faster configuration delete; deletion will be done in the background, allowing the UI to refresh immediately. Please note: this update will only be present for the Dell EMC PowerStore and Hitachi VSP integrations for IO 6.8.0. It will ship for all integrations in the January 2022 release.

Fixed bug to prevent disabling of all health notifications because of connectivity issues. There are a few properties related to this change:

Further improvements to memory use, which reduces the risk of an OOM condition in the CAPI.

More reliable restart of translators when services restart, which fixes issue where self-monitoring didn’t always restart (reconnect) when CAPI restarted.

The kill switch (system property) for ignoring health alerts on translator integrations should now function as expected.

Easier to use playback capabilities for demoing translator-based integrations.

Three new translator-based integrations: Infinidat InfiniBox, Cisco UCS, and Dell/EMC XtremIO. Be sure to review the TOI video and material for important overview and support information.

Instead of showing just the Name and Created On properties on the integration inventory page, each translator will now show Name, Tag, and five additional properties. Created On will now be hidden by default.

(Linux, Solaris, and Windows) Conversations will now be published once per minute. This schedule can be modified using the following property, which can be set on the Infrastructure Observability REST API service:

com.vi.integrations.capi.convo_cache.min_sec_between_publishes = SECONDS, where SECONDS = 60 by default.

As a result, conversations may take longer to show up the first time, but in high-churn environments, load on the graph store service should be lower.

In recent versions of CAPI, an optimization was made to base discovery that would limit non-structural changes, e.g. a tag change, to be committed only after four hours have elapsed. There was a problem with this, which has now been corrected.

As a result, the load on the graph store service should be reduced in highly dynamic environments, but at the cost of properties not updating as often. Metrics collection is not affected.

Memory use has been tuned, which reduces the risk of an OOM condition in the API when processing gets backed up.

Scope: CAPI 1.17.0 (June release of integrations).

Problem: Graph store could crash due to a “Stack overflow error”.

Resolution: We fix the cause of the stack overflow in our latest CAPI version. The Aug (6.7.7) release of core should have a more permanent fix to prevent this from happening accidentally in the future.

Scope: Upgrading from an early version of CAPI (pre 1.9.0 – Sep or earlier) to 1.16.0 (May) or later can result in a migration error that prevents CAPI from starting. This only happens if you haven’t installed a newer version of a translator AND you don’t hop to an in-between release.

Problem: CAPI fails to start due to a migration error related to the “Steal CPU” metric.

Resolution: We enhanced the migration code to check for the existence of the column before attempting migration.

Workaround: Prior to this release you could work around it by setting the com.vi.integrations.capi.skip_cpu_migration_360 property to true on the Virtual Wisdom Rest API service.

Scope:Time alignment.

Resolution: We changed time aligning to align data to the closest 30 seconds instead of the poll interval. Provides more accurate reflection of when data arrived while still providing alignment benefits in Top N reports.

The Canonical API (CAPI) has a global device tracing property that was created mainly for third-party integrations written by Virtana development partners. Currently, it makes duplicate config endpoint requests instead of just one for each configured target device. This will cause double discovery (entity) and metrics data collection, which adds unnecessary device load and other potentially unknown issues in translator-based integrations.

Please do not enable device tracing until a fix is developed and tested, likely in the July release timeframe.