Loading...
Loading...
The following issues may occur with HPE M-Series Switches running NVIDIA Cumulus version 5.13.0 (and earlier):Temporary Loss of MLAG Uplink Bond PathsIf both MLAG members have the STP mode set to PVRST and the primary member of an MLAG pair is rebooted, a temporary loss of all MLAG uplink bond paths on the secondary member immediately occurs after the primary member is back online. This results in a link disruption of approximately 5 seconds, potentially leading to an MLAG bond outage lasting up to 15 seconds.Temporary Outage by Spanning Tree ConvergenceA 3-second outage may occur due to spanning tree convergence when server and storage end nodes use auto-edge.NOTE: Perform any necessary upgrades or maintenance during a scheduled maintenance window to minimize impact.
HPE M-Series Cumulus switches in an MLAG configuration are affected by this advisory:HPE SN2010M 25GbE 18SFP28 4QSFP28 Power to Connector Airflow Half Width Switch SN2010M with NVIDIA Cumulus [S2T75A]HPE SN2100M 100GbE 16QSFP28 Power to Connector Airflow Half Width Switch SN2100M with NVIDIA Cumulus [S2T76A]HPE SN3420M 25GbE 48SFP28 12QSFP28 Power to Connector Airflow Switch SN3420M with NVIDIA Cumulus [S2T77A]HPE SN3700cM 100GbE 32QSFP28 Power to Connector Airflow Switch SN3700cM with NVIDIA Cumulus [S2T78A]HPE SN4600cM 100GbE 64QSFP28 Power to Connector Airflow Switch SN4600cM with NVIDIA Cumulus [S2T80A]HPE SN3700M 200GbE 32QSFP56 Power to Connector Airflow Switch SN3700M with NVIDIA Cumulus [S2T79A]HPE SN4700M 400GbE 32QSFPDD Power to Connector Airflow Switch SN4700M with NVIDIA Cumulus [S2T81A]
To address the temporary loss of MLAG uplink bond paths following a reboot of a switch, upgrade to HPE Cumulus version 5.14.0 (or later).Workaround for Temporary Loss of MLAG Uplink Bond PathsChange the STP protocol on the secondary switch to RSTP and then make the same change on the primary switch.For example, bond10 is the MLAG uplink bond used to connect the Cumulus switches to the core switches. Swp1 is the physical link from the server connected to the switch.Execute the following commands on the secondary Cumulus switch:nv set bridge domain br_default stp mode rstpnv set interface swp1 bridge domain br_default stp admin-edge onnv set interface bond10 bridge domain br_default stp admin-edge onnv set interface bond10 bridge domain br_default stp bpdu-guard offnv set interface bond10 bridge domain br_default stp bpdu-filter onnv config applyThen execute the same commands on the primary Cumulus switch:nv set bridge domain br_default stp mode rstpnv set interface swp1 bridge domain br_default stp admin-edge onnv set interface bond10 bridge domain br_default stp admin-edge onnv set interface bond10 bridge domain br_default stp bpdu-guard offnv set interface bond10 bridge domain br_default stp bpdu-filter onnv config applyWorkaround for Temporary Outage by Spanning Tree ConvergenceServer and storage end nodes using auto-edge should be changed to admin-edge if auto-edge spanning tree has been configured for the server and storage ports to avoid a 3-second outage by spanning tree convergence.For assistance, contact HPE Customer Support through the following methods:• Contact your local HPE Customer Support by going toHPE Services - Technical Support Phone Numbers• Log a case via theHPE Support Center
Operating Systems Affected:Not Applicable
Click on a version to see all relevant bugs
Hewlett Packard Enterprise Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.