Bug Fixes History
This section includes bug fixes history of three major releases back. For older releases history, please refer to the relevant versions.
Internal Ref. | Category | Description |
3.12.2002 | ||
3678870 | TACACS+ | Enabling logging of system changes to a TACACS+ server may cause account daemon to fail. |
4037679 | MADs | In some cases where incorrect MAD handling causes the switch to become unresponsive to the SM, a switch reset is required. |
4146212 | Credits, Data Path, Fatal Cause | Inaccurate credits configurations in switch data path which, in rare cases, can lead to unexpected behavior or performance degradation. |
4129741 | SHARP | In rare cases, when a SHARP job is stopped while ANDR operations were in use, the next job may fail and require switch reboot. |
4038484 | Certificates | Specific certificates have an incorrect expiration date displayed. |
3605188 | Port Status, QM9700 | When using an FR4-coherent optical module while running a split-unsplit flow, occasionally the port may not raise a link. |
3782281 | Optical Module, Split, QM9700 | When performing a partial split (where only one port in a cage is split) on an optical module, port may not link up. |
4035403 | SHIELDv2, pFRN | In cases of repeating flapping links, the pFRN recovery mechanism might not be triggered. |
Internal Ref. | Category | Description |
3.12.1002 | ||
3896515 | NTP | NTP settings are configured but not applied after running configuration switch-to. |
3924542 | SHARPv3 | On rare occasions, SHARP job failure with trap 135 may be experienced as a result of previous jobs that have failed or stopped. |
3811793 | Transceiver | Failure to read transceiver module information (due to bad module) may cause multiple errors and the switch may slow its response. |
3977077 | Transceiver | A flap on one port of an NDR transceiver may cause the twin port on the same transceiver to also flap. |
3813780 | Transceiver, Link Up | Following a module reset, link up may take up to 70 seconds. |
3892669 | WebUI | An issue with the web server allowed an authentication bypass, granting partial access to the switch for specific URI requests. |
3969193 | WebUI | QM9700 systems, when performing partial split port on one port in a cage, the WebUI front panel ports illustration may reflect the logical position incorrectly. |
Internal Ref. | Category | Description |
3.11.4002 | ||
3824931 3843040 | Check | Illegal packets of a permissive LID (0xFFFF) and VL other than 15 were incorrectly configured to destined for port 0 and be proceeded by the switch firmware, which lead to overloading of the switch firmware. |
3787684 | OpenSM | Customers using OpenSM on the switch may encounter switch OpenSM malfunctions. |
3864399 | pFRN | Sending pFRN packets to ports that were connected to themselves (loop), caused the switch hanged due to semaphore lock mismatch. |
3715407 | Cables, Timeout, QM9700 | Because of a race condition in cable information interfaces, the cable read information may fail. |
3860421 | Data Path, Buffer, QM9700 | Incorrect buffer configuration for trap and mirror packets may cause the switch data path to become stuck, potentially resulting in buffer overrun and internal credit leakage. |
3724050 | Debug, QM9700 | While debugging, the PRBS register could not be properly disabled which caused the port to be stuck in debug mode. |
3700332 | Credit Packet, QM9700 | Improve buffer utilization by fixing the ratio of transmitted data and credit packets on the link. |
3657798 | Management Port Split, QM9700 | During the transition from 2x split to 4x split port configuration, the procedure for clearing routing logic for management ports was incorrect. |
3707400 | SHARP, Buffer, QM9700 | In cases where buffer is overutilized, failure to allocate a buffer for the SHARP TX retry buffer may lead to buffer oversubscription. |
3608820 | SHARP, QM8700 | On rare occasions, following an event of port link down, the SHARP resources cleanup may fail. |
3858877 | TACACS+, QM8700 | Enabling logging of system changes to a TACACS+ server may cause account daemon to crash. |
3591193 3591151 3591197 | PLL Bandwidth, QM8700 | Adjusted the PLL bandwidth to accommodate certain deviations in the switch link margins for improved resilience against temperature variations. |
3.11.3006 | ||
3787192 | OpenSM | Customers running OpenSM from the switch may encounter issues where the switch stops functioning properly due to a memory leak in OpenSM. |
3793705 | Director System, Optical Cables | Director system may became slow when it is has 800 optical cables connected. |
3.11.3002 | ||
3671899 | IP Filter | Whitelisted LDAP TLS port as part of the default firewall rules. |
3733440 | Parsing | Fixed out of range parsing of port split on NDR systems. |
3.11.2016 | ||
3718528 | SHARP | In Quantum-2 switch systems, the SHARP job fails occasionally while it is running and an error event triggers a firmware dump. |
3718514 | Packet Drops | Packets sent to port0 might be dropped due to heavy load, causing MADs timeouts in SM. |
3718516 | Switch management | In rare scenarios, the user might experience OS slowness due to periodic reading of cable information and MCIA reading error in the switch logs. |
3.11.2006 | ||
3626729 | Buffers / Performance | Fixed credit management of shared buffer scheme that effected the overall bandwidth performance of the switch. |
3.11.2004 | ||
3650326 | PSPA PRM Reg / Port Status | On a switch with FR4 optical module, running split-unsplit flow or enabling L3 routing may occasionally cause the port not to raise a link. |
3468977 | IP Filter | Occasionally, some packets could pass the default IP filter rules, skipping some of them. |
3575881 | Upgrade | When updating the QM9700 OS to version 3.11.1004, an intallation wizard will appear in CLI on initial boot. |
3605465 | Reboot | In some cases, the reboot process may be interrupted and fail. |
3599629 | JSON | Show health-report can’t be parsed by json-print. |
3538336 | HDR Systems | One-time trace during leaf insertion may cause system degradation but no impact on functionality. |
3540874 | HDR Systems | One-time trace during spine insertion may cause system degradation but no impact on functionality. |
3571023 | SSH | SSH-RSA causes weak cipher. |
3.11.1014 | ||
This Release includes only FW bug fixes. | ||
35926593585886 | MVCR, Switch | Qantum-2 unmanaged switch may freeze while sending MVCR. |
358904435877033573164 | i2c, Switch | Rare issue that triggers the i2c to module connection to lock and causes the Quantum-2 switch to freeze. |
3548254 | Cables, link up | FR4 MMS4X50-NM cable link-up failure after a disconnect or AC cycle. |
3570478 | SNR | Fixed SNR value calculation for correct readings from the MMA4Z00 optical cable module. |
3311198 | Bandwidth | Disabled "low priority credits" feature on the switch side that caused the credits mechanism to overload the links with credit packets, reducing the available bandwidth for transmitting data packets on the link. |
3.11.1004 | ||
3554182 | Cables, link up | Link does not raise with 2nd source transvers MMS4X00-NS. |
3478996 | JSON | After overflooding the system with JSON requests (more than 1 million requests), system may not respond to other JSON queries. |
3538638 | Link Diagnostics | The message in the "show interfaces ib link-diagnostics" command output was incorrect. |
3509247 | Management Interfaces, IPv6 | The IPv6 route command was fixed for InfiniBand switches. |
3509292 | Management Interface, ZTP, SLAAC | SLAAC does not work well with ZTP. |
3533738 | Transceiver Diagnostics | For show interfaces transceiver command, Tx bias current shows N/A on some cables. |
3464879 | Upgrade | Upgrading LTS versions 3.10.4202 and 3.10.4300 to a 3.10.50xx and 3.10.60xx versions does not work well. The fix enables upgrade from LTS version to GA versions 3.11.10xx. |
3538638 | NDR Switch, Link Diagnostics | The message in the "show interfaces ib link-diagnostics" command output was incorrect. |
3451519 | ibdiagnet, Module Temperature Alarm Type | When using ibdiagnet, an incorrect module alarm type was reported. |
3448282 | HDR Switch, BER | On HDR switches, at times, the link gets stuck due to unresponsive peer having very high BER. |
3447029 | HDR Switches, Port, Reboot, BER | On HDR switches, on some occasions, if a peer's port resets or is physically toggled, the port may be reported as unhealthy due to BER threshold crossing. |
3407038 | I2C | An unresponsive PSU client can cause the SDA I2C line to hang. |
3499997 | NDR Switch, SHARP | In some cases, the combination of SHARP SAT traffic and SHARP MADs can cause the switch to get stuck. |
3.10.6004 | ||
3434543 | LDAP SSL | LDAP SSL option (ldap ssl mode ssl) is missing from the CLI. For more information see ldap sll command in "User Management and Security Commands" section in the user manual. |
3354452 | NDR Router, Multi-SWID | In QM9700 systems, in multi-SWID profile, the following CLI command "no sma port <subnet number> shutdown" will not work properly. |
3355335 | NDR Router, Multi-SWID | For QM9700 systems, on multi-SWID profile, when configuring through an external openSM a subnet-prefix with the value "0xfec0000000000000", the subnet state shown in CLI may appear as down instead of UP. |
3335057 | Port LED, Optic Cables | In QM9700 systems, when a transceiver module is plugged in when only one of the optic cables is connected (while the second cable is disconnected), the port LED may be incorrectly displayed on the disconnected side. |
2824249 | Installation, Firmware | After a firmware update failure, the bad image was not erased. |
3283303/ 3298590 | Link Maintenance | In the rare event of an error burst, the link maintenance stopped working. |
3339363 | pFRN | pFRN notification state machine got halted in busy-wait on all riscs due to inability to free TX credits. |
3342918 | Port Link Up, Port Toggling, Optical Modules | On rare occasions, the port might get stuck (in all speeds) during the link up flow when using optical modules. |
3336788 | SHARP | When operating in dynamic trees allocation mode, MAD error responses might be received in libsharp. |
3436317 | SHARP | In HDR systems, on rare occasions, when a SHARP QP exceeds the allowed amount of retries, the switch may hang due to an incorrect flow execution. |
3389432 | SM, Timeout, Flint, Failure | In QM9700 systems, the flint burning firmware process might take longer than expected, possibly leading to timeouts in SM and logical links drops by the SM, which, in turn, may lead to failure of the flint burn command. |
3362200 | Turbo Path | In rare cases that involve stress of traffic, unexpected hardware fast path behavior may occur, possibly leading to the switch firmware hanging when toggling the ports. |
3.10.5002 | ||
3362685 | Port LED, Optic Cables | In QM9700 systems, when a transceiver module is plugged in when only one of the optic cables is connected (while the second cable is disconnected), the port LED may be incorrectly displayed on the disconnected side. |
3377608 | sharp_am, libsharp | When operating in dynamic trees allocation mode, MAD error responses might be received in libsharp. |
3362200 | Turbo Path | In rare cases that involve stress of traffic, unexpected hardware fast path behavior may occur, possibly leading to the switch firmware hanging when toggling the ports. |
3.10.5000 | ||
3264093 | CS7500, BER | On CS7500 systems, in cases of high error rate on one of ports connected, the firmware may hang, causing the module to go into fatal mode. |
3301825 | Counters | The firmware does not return values for the counters "PortSwLifetimeLimitDiscards" and "PortSwHOQLifetimeLimitDiscards". Support has now been added for the counters. |
3230314 | Health Notification | For 1U systems, in cases where the MGMT module becomes fatal, a health alert is not sent properly. |
3291564 | Logging | The following log is printed once per minute (instead of not printing at all): “bind(23) AF_INET6 ::1#123 flags 0×5 failed: Cannot assign requested address” |
3335002 | PFRN | pFRN mirror v1 header pad count showed an invalid padding size. |
3261861 | Quantum, Quantum-2, HDR-to-NDR | Connecting an HDR device to an NDR device with Optical cables longer than 30m causes degradation in the bandwidth. |
2974424 | Quantum-2, Cables, Polarity Inversion | Currently, there is no link up when connecting to Quantum-2 systems with cables that perform polarity inversion. |
3332824 | Quantum-2, LED | On QM9700, when any fan is plugged out, the front panel fan status LED remains green instead of turning amber. |
3331691 | Quantum-2, LED | On QM9700, when FAN7 is plugged out, the rear panel FAN7 status LED remains green instead turning amber. |
3335044 | Quantum-2, Ports, Display | On QM9700, on the WebUI "Ports" page, the port cages that are located on the bottom row of front panel had an incorrect visual display for sub-ports 1 and 2. |
3143685 | Quantum-2, Transceiver | The switch does not return SN or PN when trying to call via mlxlink or ibdiagnet. |
3438507 | SHARP | On rare occasions, on switches with SHARP configuration, the ASIC may hang due to an incorrect flow execution. |
Internal Ref. | Category | Description |
3.10.4202 | ||
3364526 | Boot, I2C | On are occasion, a race condition occurred in the platform driver that causes I2C bus initialization failure during system boot. |
3264093 | CS7500, BER | On CS7500 systems, in cases of high error rate on one of ports connected, the firmware may hang, causing the module to go into fatal mode. |
3375327 | CS8500, LID | On CS8500 systems, when changing locally attached device, a bad LID may be reported. |
3261861 | HDR-to-NDR | Connecting an HDR device to an NDR device with Optical cables longer than 30m causes degradation in the bandwidth. |
3315167 | NTP, Logging | When NTP is configured, the following print may appear to the log every minute:
|
3364639 | Quantum, Quantum-2, pFRN | When pFRN is configured, on very rare occasions, when more than nine ports go down at once, the ASIC may not respond and may not run traffic properly. |
3372998 | Quantum-2, sharp_am, libsharp | MAD error responses might be received in libsharp when operating in dynamic trees allocation mode. |
3232534 | Switch-IB 2, Link, Port | On some occasions, when shutting down a port, the link can take a few additional seconds to close the port. |
3335044 | Web UI, QM9700, Ports | On QM9700, on the WebUI "Ports" page, the port cages that are located on the bottom row of front panel had an incorrect visual display for sub-ports 1 and 2. |
Internal Ref. | Category | Description |
3.10.4100 | ||
3236568 | Cables, CMIS | The CMIS cables lengths were rounded. Improved accuracy of CMIS cables length to display exact cable length. |
3284414 | EULA | On WebUI and CLI, the EULA document did not appear properly. |
3261861 | HDR-to-NDR | Connecting an HDR device to an NDR device with Optical cables longer than 30m causes degradation in the bandwidth. |
3283539 | Power Supply, Fan | When power supply cable is removed and inserted again, the fan speed sometimes did not set properly on the power supply fan. |
3269531 | Quantum, MSPS | On Quantum switches, after multiple MSPS (Management System Power Supply register) calls, the switch gets stuck. |
3441229 | Quantum, SHARP | On rare occasions, on switches with SHARP configuration, the ASIC may hang due to an incorrect flow execution. |
3199650 | Quantum, Quantum-2, SHARP | A physical link failure between switches while a SHARP job is running and utilizing the link can cause one of the switches to become invalid for further SHARP jobs. This can result in either a "No resource" response for new SHARP job requests or in jobs getting stuck. The bug fix requires SHARP version 3.2. |
3267152 | Quantum-2, BER COLLECT | On NDR devices, when collecting BER data, the peer falls, causing the switch to hang. |
3253800 | Switch-IB 2, CPLD | SB78xx systems show a boot error message when hw-mgmt tried to read the CPLD version, even though correct CPLD versions were reported for CPLD1 and CPLD2. |
3.10.4006 | ||
3223270 | CS8500, LIDR | On CS8500 systems spine modules, in some cases, timeout on applications that send LIDR MAD packets (such as smpquery) may occur. |
3173334 | Management Interfaces | On systems that include two management interfaces, when connecting mgmt1, ping may not be available for mgmt0. |
3254801 | Quantum, QM8700, Chassis Management | On QM8700 systems with version 3.10.3100, some of the voltage sensors may indicate wrong values which could result in a false health alert and an incorrect CLI output. |
3245821 | Quantum, Quantum-2, PFRN | In case of an AR group table set request, the ARN mask is flushed for group that has an active pFRN timer. |
3242209 | Quantum, Quantum-2, PFRN | Set PFRN mad did not return error on wrong inputs in mask_clear_timer and mask_force_clear_timer fields. |
3253717 | Quantum, Quantum-2, pFRN | mask_force_clear_timeout timer in pFRN feature was not functional (the mask was not cleared when the timer expired). |
3.10.3100 | ||
3172779 | ACL, IPv6 | The ACLD may crash when ACL with log action captures an IPv6 frame with a very large number of nested headers (total header length larger then 512B). Unexpected system behavior may occur due to the crash. |
3172851 | JSON API Authentication | In version 3.10.300x, JSON API authentication fails when posting to /admin/launch?script=rh&template=json-request&action=json-login. |
3190552 | Secure Upgrade | On QM9700 systems, a secure upgrade from 3.10.300x to a higher version depends on an updated system date. If the date is not in the range of September 1, 2021 to June 1, 2036, the upgrade will not work. |
3135787 | Sysdump | On rare occasions, during the snapshot creation, sysdump was not automatically generated. |
3.10.3002 | ||
3096687 | CS85xx, CS75xx, Subnet Manager | During a standby switch management failure, the spine of the switch may be incorrectly configured, possibly causing it to have the same LID as another spine. |
2878837 | InfiniBand Interface, Ports | If the end subport is not specified, the current behavior of "interface ib <start_port-<end_port> " will change to include all subports of the end port in the command (e.g., interface ib 1/1-1/3 shutdown will shut down ports 1/1,1/2,1/3 or, in case of split, 1/1,1/2,1/3/1,1/3/2). |
3085287 | Password Hardening | On some occasions, after setting a password for new user, an internal error in password hardening may appear. |
2998597 | Quantum, Performance | In Quantum systems, bandwidth degradation may be visible in large scale random traffic patterns (e.g., all2all and Adaptive Routing) due to wrong fast path configurations. |
2974424 | Quantum-2, Cables, Polarity Inversion | Currently, there is no link up when connecting to Quantum-2 systems with cables that perform polarity inversion. |
3082569 | Quantum-2, Counters | In some traffic patterns involving small packets, the PortRcvErrors counter may mistakenly count events of local physical errors due to an internal flow in the hardware that involves link packets. |
3085427 | Quantum-2, SHARPv3 | On rare occasions, SHARP semaphore may remain locked on a port following an event of a port link down or an application crash. |
3011581 | Quantum-2, SHARPv3 | On rare occasions, job failures with SharpError trap may be experienced as a result of previous jobs that have failed. |
3127727 | Quantum-2, Switch Hang, Fast Path, Split | On rare occasion, on Quantum-2 systems, when egress port is split to two, the egress port may get stuck due to wrong Fast Path configuration. |
3129809 | WebUI | The default certificates (rather than customer-loaded certificates) were used upon conducting a reload. |
3.10.2202 | ||
3000602 | Cables | After disconnecting MMS4X00-NL* cable and connecting Ultron cable to the same port, the port fails to link up. |
3060122 | SHARPv3 | In the event of link fault of a link between root switch and non-root switch during the run of a job, the next job run on the non-root switch may fail. |
3077375 | Upgrade, IP Filter | When upgrading to 3.10.20xx with existing IP filter configuration, the upgrade will fail and there will be a rollback to the older software version. |
3.10.2102 | ||
3040232 | PLFT, SMA | PLFT mapping for SMA port (port 0) was configured in a way that caused PLFT of FDB 0 to be used instead of PLFT of FDB 1. |
2859363 | Quantum-2, Auto-Negotiation | When using Quantum-2 systems in Auto-Neg mode, NDR speed in one lane (1x) is not supported. |
3033131 | Qunatum-2, SHARPv3 | The number of flows changed from 2 to 1, as intended. |
3003011 | Memory | A memory leak in the master module caused it to stall, causing the standby module to take over and become master. |
3.10.2000 | ||
2950888 | Certification | When whitespaces appeared in the "flag" certificate, caused certificate parsing failure. |
2963304 | Email Notifications | White spaces in the "flag" certificate caused certificate parsing failure. |
2918315 | OpenSM | Tools that require MAD RMPP protocol (such as saquery and ibqueryerrors) fail with "Connection timed out" error due to an SDK bug that prevented RMPP SM responses from being caught. |
2982113 | SHARPv3 | On rare occasions, job resource cleanup may fail. |
2972388 | SHARPv3 | Running of concurrent jobs may lead to states where jobs unexpectedly terminate or get stuck. |
2971339 | SHARPv3 | Users may experience performance degradation on high load scenarios, |
3004309 | SSH | Applying "ssh server listen interface [interface name]" caused improper SSHD configuration, preventing the establishment of a new SSH session. Sessions established before mentioned configuration remained. |
3.10.1214 | ||
2910161 | Quantum-2, Auto-Negotiation, Copper Cables | In auto-negotiation flow, using copper cables when toggling both port's sides may cause the port to get stuck on rare occasions. |
2849215 | Quantum-2, Cables | On Quantum-2 switches, when working with MFA7U10-H0xx cables, if one of the ports in a cage is disabled at the time of initialization by user configuration, reenabling the port will require toggling the link (i.e., enable → disable → enable). |
2885798 | Quantum-2, Copper Cables | In Quantum-2 systems, effective errors may occur with short Copper cable MCP4Y10-N00B. |
2890632 | Quantum-2, Optical Modules | Currently, on Quantum-2 systems, changing the Optical module rate is not allowed. |
2933388 | Quantum, CS8500, Speeds | In CS8500 platform, the system may get stuck when trying to raise a link in an unsupported speed when using Optical cables. |
2709851 | SM Traps | In some cases, traps that were sent when there is a change in link state may not be sent to SM due to wrong logic of the link state machine in the firmware. |
3.10.1202 | ||
2833052 | Password | Changing the user password and then proceeding SSH login with the new password affects configurations. When running "show configuration files", the "unsaved changes" field is reflected as "yes" (instead of "no"). |
3.10.1110 | ||
2863063 | Quantum-2, Split | Split is currently not supported in Quantum-2 systems. |
2798848 | Quantum-2, Tx | In Quantum-2 systems, changing the SerDes Tx set values is not operational. |
3.10.1000/1016 | ||
2798556 | BER | Modified the Effective BER calculation method. Due to this, the value of the Effective BER will be slightly higher, however the link quality remains the same as prior to this change. Note: There is no change in the Symbol BER, it is the same as prior to this change. |
2803887 | Logging | Updated format of port state change event print to log to include severity. |
2836940 | Quantum-2, InfiniBand Data Packet | InfiniBand traffic is not passing with operational vl 1 configured on switch interfaces. |
2837883 | Quantum-2, Link Up | Using Quantum-2 systems with copper cables and toggling all ports frequently, on rare occasions, may cause a port may get stuck and no longer link up. |
2853051 | Recovery | For CS8500 systems, when spine at slot 1 gets into error state, errors may occur during recovery flow. |