News
Abstract
It was observed that during partition migration or during recovery of a virtual Network Interface Controller (vNIC) device after failover, certain sequences of events can occur that might cause network connection failures.
Content
Linux Releases Affected
Red Hat® Enterprise Linux® 8.1
Red Hat Enterprise Linux 8.2
Red Hat Enterprise Linux 8.3
SUSE Linux Enterprise Server 15, Service Pack 1
SUSE Linux Enterprise Server 15, Service Pack 2
Red Hat® Enterprise Linux® 8.1
Red Hat Enterprise Linux 8.2
Red Hat Enterprise Linux 8.3
SUSE Linux Enterprise Server 15, Service Pack 1
SUSE Linux Enterprise Server 15, Service Pack 2
SUSE Linux Enterprise Server 15, Service Pack 3
IBM Systems Affected
All IBM® Power Systems that support vNIC.
All IBM® Power Systems that support vNIC.
Description
Partition migration or recovery of a vNIC device after a failover might leave one or more network interfaces in an inoperable state and might result in the following conditions:
• Intermittent loss of network connectivity.
• Loss of network connectivity, that requires maintenance.
• Loss of Resource Monitoring and Control (RMC) connection with the Hardware Management Console (HMC) that requires maintenance.
• Intermittent loss of network connectivity.
• Loss of network connectivity, that requires maintenance.
• Loss of Resource Monitoring and Control (RMC) connection with the Hardware Management Console (HMC) that requires maintenance.
To recover the system, perform the following steps:
Note: It might be necessary to restart the link state to reestablish the connection.
1. To list all network interfaces on the LPAR, run the following command:
Note: It might be necessary to restart the link state to reestablish the connection.
1. To list all network interfaces on the LPAR, run the following command:
# ip link show
2. To establish the link state, run the following commands:
2. To establish the link state, run the following commands:
# ip link set <interfaceX> down
# ip link set <interfaceX> up
3. In some cases, it might be necessary to unbind and rebind the interface to the driver. To unbind and rebind the interface to the driver, run the following commands:
# ip link set <interfaceX> up
3. In some cases, it might be necessary to unbind and rebind the interface to the driver. To unbind and rebind the interface to the driver, run the following commands:
# ofpathname envX | awk -F@ '{print $2}'
3000000X
# echo 3000000X > /sys/bus/vio/drivers/ibmvnic/unbind
# echo 3000000X > /sys/bus/vio/drivers/ibmvnic/bind
3000000X
# echo 3000000X > /sys/bus/vio/drivers/ibmvnic/unbind
# echo 3000000X > /sys/bus/vio/drivers/ibmvnic/bind
Workaround
• You can use alternatives to vNIC failover, such as active-backup bonding of a vNIC and a virtual Ethernet (veth), to reduce exposure to failover.
• Avoid performing Live Partition Mobility (LPM) under heavy load to reduce exposure to LPM defects.
• Avoid performing Live Partition Mobility (LPM) under heavy load to reduce exposure to LPM defects.
Fix Outlook
IBM is working closely with SUSE and Red Hat to release a fix for this issue. The fix should come as part of a future SLES or RHEL maintenance release. Please open a support ticket with SUSE or Red Hat if a hotfix is needed prior to the next maintenance release.
See Red Hat Bug: 1913324
See SUSE Bug: 1180612, 1186206, 1192273
See SUSE Bug: 1180612, 1186206, 1192273
[{"Type":"MASTER","Line of Business":{"code":"","label":""},"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"HW1W1","label":"Power -\u003EPowerLinux"},"ARM Category":[],"ARM Case Number":[],"Platform":[{"code":"PF025","label":"Platform Independent"}]}]
Product Synonym
Red Hat Enterprise Linux 8.1, or later
SUSE Linux Enterprise Server 15, Service Pack 1, or later
Was this topic helpful?
Document Information
Modified date:
02 December 2021
UID
ibm16479855