OS becomes unresponsive when Intel E810 NICs are added to a bond
This document (000020969) is provided subject to the disclaimer at the end of this document.
Environment
SUSE Linux Enterprise Server 15 SP4
Situation
A - When Intel E810 device NIC ports are added to a bond device, the OS hang may be observed in following scenarios:
1. System is rebooted
2. A slave interface is removed from the bond device
3. A netdev_unregister event is generated for any other NIC in the system
B - Hangs, enabling bond, were reported again, starting with kernel 5.14.21-150400.24.41.
The system will hang on reboot, restarting network services after bond was created.
The system will also hang running any network related commands.
1. System is rebooted
2. A slave interface is removed from the bond device
3. A netdev_unregister event is generated for any other NIC in the system
B - Hangs, enabling bond, were reported again, starting with kernel 5.14.21-150400.24.41.
The system will hang on reboot, restarting network services after bond was created.
The system will also hang running any network related commands.
Resolution
A - The issue is resolved in the SLES15 SP4 kernel update, November/2022, kernel 5.14.21-150400.24.33 and kernel 5.14.21-150400.24.38.1.
Workaround:
A work around is not available if bonding was already created.
If the OS becomes hung, system may have to be reset.
Once reset, the following steps will help to avoid the issue on next reboot:
1. Boot into single user mode
2. cd to /etc/sysconfig/network-scripts/
3. Delete the ifcfg-* files related to bonding.
4. Boot into multi user mode
5. Install maintenance update kernel version: 5.14.21-150400.24.33 or later.
B - For the second scenario observed with kernel 5.14.21-150400.24.41 and later:
The kernel update released May/2023, kernel 5.14.21-150400.24.63.1 resolves the issue.
Alternatively the following workaround (blacklisting irdma) can be used:
Workaround:
A work around is not available if bonding was already created.
If the OS becomes hung, system may have to be reset.
Once reset, the following steps will help to avoid the issue on next reboot:
1. Boot into single user mode
2. cd to /etc/sysconfig/network-scripts/
3. Delete the ifcfg-* files related to bonding.
4. Boot into multi user mode
5. Install maintenance update kernel version: 5.14.21-150400.24.33 or later.
B - For the second scenario observed with kernel 5.14.21-150400.24.41 and later:
The kernel update released May/2023, kernel 5.14.21-150400.24.63.1 resolves the issue.
Alternatively the following workaround (blacklisting irdma) can be used:
modprobe -r irdma modprobe -r ice echo 'blacklist irdma' >> /etc/modprobe.d/sfsc00603511.conf echo 'install irdma /bin/false' >> /etc/modprobe.d/sfsc00603511.conf modprobe ice
Disclaimer
This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.
- Document ID:000020969
- Creation Date: 08-Feb-2023
- Modified Date:12-May-2023
-
- SUSE Linux Enterprise Server
For questions or concerns with the SUSE Knowledgebase please contact: tidfeedback[at]suse.com
