Bug Fixes History

NVIDIA UFM-SDN Appliance User Manual v4.16.0

Ref. #

Description

Rev 4.15.0

3752196

Description: Fixed intermittent UFM REST API Failures

Keywords: REST API, Failure

Discovered in Release: 4.14.1

3747086

Description: Calling sh ib managed-switch auto-provisioning status causes error

Keywords: Fixed managed-switch, auto-provisioning

Discovered in Release: 4.14.0

Rev 4.14.1

3670183

Description: Monitoring endpoint not returning counters for an active interface

Keywords: Monitoring, Active Interface, Counters

Discovered in release: v4.14..0

3670182

Description: Inconsistent port format type returned from the UFM

Keywords: Inconsistent, Port, Format Type

Discovered in release: v4.13.1

3666944

Description: Port auto isolation failed to activate when a port consistently exhibited a high Symbol BER (1e-7)

Keywords: Port Auto Isolation, Symbol BER

Discovered in release: v4.13.1

3665316

Description: The UFM REST API endpoint /ufmRest/resources/ports provide inaccurate port state information

Keywords: Ports REST API, Port State

Discovered in release: v4.13.1

3604194

Description: UFM Fabric Validation "CheckPortCounters" failure

Keywords: Fabric Validation, CheckPortCounters

Discovered in release: v4.12.2

Rev 4.14.0

3644553

Description: When querying the ports, adding a cable_info=true as an argument will give cable information per port

Keywords: Ports, Query, cable_info=true

Discovered in release: v4.13.0

3604212

Description: Broken UFM REST API

Keywords: REST API

Discovered in release: v4.12.1

3604183

Description: UFM error interpretation

Keywords: Error, Interpretation

Discovered in release: v4.13.1

3587849

Description: OpenSM restart when backup UFM loses power

Keywords: OpenSM, Restart

Discovered in release: v4.8.0

3577427

Description: UFM REST API returns wrong model for NDR unmanaged switch

Keywords: Unmanaged Switch, NDR, REST API

Discovered in release: v4.12.1

3575882

Description: UFM event is not generated for a switch down

Keywords: UFM Event, Switch Down

Discovered in release: v4.12.1

3628421

Description: UFM timezone issue

Keywords: Timezone

Discovered in release: v4.13.1

Rev 4.13.1

3590777

Description: After upgrading UFM new telemetry data is not being collected and presented in UI Telemetry tab.

Keywords: Telemetry, Coredump

Discovered in release: 4.13.0

3461058

Description: When using the Dynamic Telemetry API to create a new telemetry instance, the log rotation mechanism will not be applied for the newly generated logs of the UFM Telemetry instance

Keywords: Dynamic, Telemetry, Log-rotate

Discovered in Release: 4.12.0

3383916

Description: Fixed Client CTRL+C server disruption

Keywords: Client CTRL+C, Server functionality

Discovered in Release: Rest Over RDMA Image 1.0.0-21

3375414

Description: Fixed improper functionality of UFM UI Dashboard

Keywords: UI Dashboard

Discovered in Release: 4.10.0

3342713

Description: Fixed UFM Health configuration for periodic restarts of the telemetry

Keywords: UFM Health, Telemetry, Periodic restarts

Discovered in Release: 4.10.0

3457714

Description: The Subnet Merger function in the NDT plugin is not supported in UFM Appliance.

Keywords: UFM Subnet Merger; Plugin

Discovered in Release: 4.11.0

3361160

Description: Fixed UFM long upgrade time due to a large historical Telemetry database file

Keywords: Long Upgrade Time, Historical Telemetry, Database File

Discovered in Release: 4.10.0

3268270

Description: Show managed switches inventory data (Sysinfo) immediately after UFM initialization

Keywords: Managed Switches, Inventory, Sysinfo

Discovered in Release: 4.10.0

3338600

Description: Fixed UFM UI lockdown by adding protection to the failed path on backend side

Keywords: UFM UI, lockdown

Discovered in Release: 4.10.0

3276163

Description: Fixed remote syslog configuration in UFM Web UI to be persistent

Keywords: Remote Syslog, Web UI

Discovered in Release: 4.10.0

3268270

Description: Show managed switches inventory data (Sysinfo) immediately after UFM initialization

Keywords: Managed Switches, Inventory, Sysinfo

Discovered in Release: 4.10.0

3338600

Description: Fixed UFM UI lockdown by adding protection to the failed path on backend side

Keywords: UFM UI, lockdown

Discovered in Release: 4.10.0

3276163

Description: Fixed remote syslog configuration in UFM Web UI to be persistent

Keywords: Remote Syslog, Web UI

Discovered in Release: 4.10.0

3266866

Description: Failure to collect system dump for unmanaged switch

Keywords: System Dump; Unmanaged Switch

Discovered in Release: 4.10.0

3234082

Description: UFM WebUI unresponsive after failover issue

Keywords: UFM, WebUI, failover

Discovered in Release: 4.9.0

3199572

Description: Incorrect Tier reporting in the UFM events

Keywords: Tier, Incorrect Report

Discovered in Release: 4.9.0

3107006

Description: Using GET All Modules REST API (GET /ufmRest/resources/modules), returns N/A in device_name.

Keywords: Modules, N/A, device_name

Discovered in Release: 4.8

3076817

Description: Upgrading to the latest UFM version (UFMAPL_4.8.0.6_UFM_6.9.0.7), the UFM WEB UI shows log and error messages with "invalid date."

Keywords: WEB UI, "invalid date"

Discovered in Release: 4.8

3060127

Description: UFM WEB UI - Ports REST API returns tier parameters as N/A in response

Keywords: WEB UI, tier, N/A

Discovered in Release: 4.8

3052660

Description: UFM monitoring mode is not working

Keywords: Monitoring, mode

Discovered in Release: 4.8

3031121

Description: Network map showing a link between QM8790 and Manta Ray leaf having BW of >20,000 Gb/s​

Keywords: Network Map, BW, 20,000

Discovered in release: 4.7.0

3003366

Description: UFM Starting and Stopping On Its Own Since Merge​

Keywords: Start, Stop

Discovered in release: 4.6.0

2968236

Description: Fabric health Old Alerts and events do not clear

Keywords: Fabric Health, Alerts, clear

Discovered in release: 4.7.0

2957984

Description: BER Not Being Read or Reported​

Keywords: BER, Not, Reported

Discovered in release: 4.7.0

3032227

Description: UFM UFMAPL_4.7.0.3_UFM_6.8.0.6 lists one of my skyways as "host" instead of "gateway"​

Keywords: skyway, gateway, host

Discovered in release: 4.7.0

2966472

Description: UFM Fabric health BER_CHECK warnings​

Keywords: Fabric Health, BER, check

Discovered in release: 4.7.0

2968236

Description: All ibdiagnet runs fail if multi-port SM is enabled

Keywords: Multi-port SM, Fabric Health report, ibdiagnet

Fixed in release: 4.7.1

2972007

Description: UFM cannot start if peer machine is not reachable

Keywords: High availability, UFM start

Fixed in release: 4.7.1

2957984

Description: High BER ports are not being read or reported.

Keywords: High BER ports

Fixed in Release: 4.7.1

-

Description: UFM does not automatically isolate high Effective BER ports.

Keywords: Unhealthy ports, auto-isolated high BER ports

Fixed in release: 4.7.0

2705974

Description: If new ports are added after UFM startup, the default session REST API (GET /ufmRest/monitoring/session/0/data) will not include port statistics for the newly added ports.

Keywords: Default session; REST API; missing ports

Fixed in release: 4.7.0

2804807

Description: UFM WEB GUI becomes unresponsive and event/REST API log stops printing.

Keywords: Web UI, unresponsive

Fixed in release: 4.7.0

2801258

Description: UFM failed to serve incoming REST API requests.

Keywords: REST API, hang, unresponsive

Fixed in release: 4.7.0

2744757

Description: CheckSMConnectivityOnStandby should consider multiple GUIDs on a port.

Keywords: UFM Health, Multi-port SM

Fixed in release: 4.6.0

2589617

Description: UFM stopped to discover topology on SuperPOD environment.

Keywords: Stopped; discover

Fixed in release: 4.6.0

2615514

Description: UFM software no longer supports license type "UFM APPLIANCE".

Keywords: License; UFM Appliance

Fixed in release: 4.6.0

2677501

Description: UFM triggers "Cooling Device Monitoring Error" event when AHX monitoring is configured.

Keywords: AHX monitoring

Fixed in release: 4.6.0

2790127

Description: UFM Fabric Health report aborted with a timeout.

Keywords: Fabric Health report, ibdiagnet

Fixed in release: 4.6.0

2789146

The CLI command "ib topo-file generate" not working.

Fabric Topology file

Fixed in release: 4.6.0

2782069

Description: (None) values are reported in "High BER port" UFM events.

Keywords: High BER Port

Fixed in release: 4.6.0

2482401

Description: Telemetry is not reporting multi-HCA ports due to naming convention

Keywords: Telemetry: multi-HCA

Fixed in release 4.5.1

2638032

Description: Wrong label in "Effective BER" event.

Keywords: Effective BER; events

Fixed in release 4.5.1

2288773

Description: NVIDIA-SuperPOD UFM 6.5.0-4 incorrectly reports HCA names in GUI

Keywords: WebUI

Discovered in Release: 4.3.0

Fixed in Release: 4.4.0

1967348

Description: Email sender address cannot contain more than one period (“.”) in the domain name.

Keywords: Email, sender, period

Discovered in Release: 4.1.5

Fixed in Release: 4.2

2069425

Description: SMTP server username cannot have more than 20 characters.

Keywords: Email

Discovered in Release: 4.1.5

Fixed in Release: 4.2

2197199

Description: Auto-provisioning cannot communicate with Mellanox switch over HTTPS.

Keywords: Auto-provisioning

Discovered in Release: 4.1.5

Fixed in Release: 4.2

2192488

Description: FW upgrade as part of UFM Appliance upgrade causes to QP reset fail.

Keywords: UFM Appliance upgrade

Discovered in Release: 4.1.5

Fixed in Release: 4.2

1783048

Description: Opening UFM web UI in monitoring mode is now supported.

Keywords: Web UI, monitoring mode

Discovered in Release: 4.1.5

Fixed in Release: 4.2

1691882

Description: UFM Agent now is now part of the UFM web UI.

Keywords: UFM Agent

Discovered in Release: 4.1.5

Fixed in Release: 4.2

1793244

Description: UFM/module temperature thresholds notifications.

Keywords: Temperature module

Discovered in Release: 4.1.5

Fixed in Release: 4.2

1358248

Description: Fixed the issue where ibdiagnet’s unresponsiveness when using the get_physical_info flag caused UFM to hang.

Keywords: ibdiagnet

Discovered in Release: 3.7.0

Fixed in Release: 4.1.5

1706226

Description: Fixed an issue where MLNX_OS credentials were missing at the device "access_credentials" menu (the issue was detected on old Java based GUI). At the new UFM Web UI – MLNX_OS credentials are represented by HTTP credentials.

Keywords: MLNX_OS, credentials

Discovered in Release: 4.0.0

Fixed in Release: 4.1.5

1914379

Description: MellanoxCare service can now communicate with UFM (valid only when http communication is configured between MCare and UFM).

Keywords: MellanoxCare, http, https

Discovered in Release: 4.0.0

Fixed in Release: 4.1.5

1779694

Description: Fixed the issue where UFM REST API "get_all_pkeys" broke backward compatibility.

Keywords: UFM REST API, PKey Management

Discovered in Release: 4.0.0

Fixed in Release: 4.0.1

1450154

Description: Fixed the issue where partition configuration was lost after UFM restart.

Keywords: partitions.conf, PKey, configuration

Discovered in Release: 3.5.0

Fixed in Release: 3.7.2

1498857

Description: Fixed the issue where the SM used to crash during tear down when SM port GUID is 0.

Keywords: Subnet Manager

Discovered in Release: 3.5.0

Fixed in Release: 3.7.2

1422797

Description: Added missing width_enabled and width_supported attributes to the ports REST APIs.

Keywords: REST

Discovered in Release: 3.5.0

Fixed in Release: 3.7.2

1294010

Description: Fixed the issue where partition configuration was lost after upgrading to UFM version 5.9.6 and restarting the server.

Keywords: partitions.conf, PKey, configuration

Discovered in Release: 3.6.0

Fixed in Release: 3.7.0

1276539

Description: Updated report execution command in order to avoid the following false warning of wrong link speed during topology comparison:

"Wrong link speed on:Sf4521403000f0c10/Nf4521403000f0c10/P1. Expected:FDR10 got:10"

Keywords: Topology compare report

Discovered in Release: 3.6.0

Fixed in Release: 3.7.0

1131286

Description: Fixed a memory leak of UFM’s main process when running multiple reports periodically.

Keywords: Memory leak, reports

Discovered in Release: 3.5

Fixed in Release: 3.6.1

1064349

Description: Fixed the issue where UFM health reported false unresponsiveness of OpenSM.

Keywords: UFM health, OpenSM

Discovered in Release: 3.0

Fixed in Release: 3.5

920493

Description: Fixed the issue where UFM health used to crash when restarting history mysql daemon.

Keywords: UFM health, history mysql

Discovered in Release: 2.5

Fixed in Release: 3.0

958428

Description: Fixed the issue where UFM fabric health report used to crash when "Cable type check & cable diagnostics" was checked in settings.

Keywords: UFM fabric health reports

Discovered in Release: 2.5

Fixed in Release: 3.0

798876

Description: Fixed the issue where there was no indication of removing an electric cable out of a power supply module.

Keywords: UFM events, device management

Discovered in Release: 2.5

Fixed in Release: 3.0

-

Description: Fixed the issue of when one of the ports of ConnectX-4 card was removed and the other remained connected, a switch module removal event was created.

Keywords: UFM events, Device Management

Discovered in Release: 2.5

Fixed in Release: 3.0

-

Description: Fixed the issue where UFM CPU usage test was greater than the threshold (100%).

Keywords: UFM CPU

Discovered in Release: 2.5

Fixed in Release: 3.0

688875

Description: Fixed the invalid format (Grid [Grid [Grid]]) of the Event Logs.

Keywords: UFM events log

Discovered in Release: 2.5

Fixed in Release: 3.0

-

Description: Fixed the issue where UFM GUI launch used to fail due to Java security block.

Keywords: UFM GUI

Discovered in Release: 2.5

Fixed in Release: 3.0

-

Description: Fixed the issue where Mellanox Care failed to complete scan (IB Simulator failure).

Keywords: Mellanox Care

Discovered in Release: 2.5

Fixed in Release: 3.0

766828

Description: Fixed the issue of a false "SM LID Changed" event was created in monitoring mode.

Keywords: UFM events, Monitoring mode

Discovered in Release: 2.5

Fixed in Release: 3.0

690862

Description: Fixed the issue where ConnectX-4 grouping on UFM was corrupted when using the Unhealthy Ports feature.

Keywords: Grouping, unhealthy, ports

Discovered in Release: 2.5

Fixed in Release: 3.0

© Copyright 2024, NVIDIA. Last updated on May 24, 2024.