< back to search

gpp_maybe Recommended update for suse-migration-services
SUSE-SLES-16.0-1282

This update for suse-migration-services fixes the following issues: Changes in suse-migration-services: - yq is only recommended not required - Bump version: 2.1.34 → 2.1.35 - Update suse-migration-rpm - Handle multipath root (#493) Add handling of multipath_wnn kernel parameter for multipath root. This allows to detect multipath mapped systems and mount them properly in the DMS live migration system and Fixes #491 - Explicitly request python3-base - Make sure to add multipath to the initrd for SAP (bsc#1260340) - Fix install of NetworkManager - Fix missing tests for latest apparmor code change - Bind mount resolv.conf to /system-root/run/.../resolv.conf - Mount /system-root/run as tmpfs Use tmpfs for /run, so we do not write temporal files during migration to disk. - Make sure to add multipath to the initrd (bsc#1260340) - Fix loading SELinux policy of Salt Bundle with workaround - preserve resolv.conf (bsc#1263889) - Avoid temporary test/unit/.coverage in dist archive This temporary file, generated by pytest-cov, is created within each `make test` call containing different content, also if no source files has changed. - Integrate ResolvConf class and reorder network setup services Refactor SetupNameResolver and WickedToNetworkManager to use the new ResolvConf class for unified resolver handling. Swap systemd service dependencies to ensure name resolver setup runs before host network setup. - Add ResolvConf class for unified name resolver mgmt Implement a new ResolvConf class to provide a consistent approach for handling /etc/resolv.conf during migrations. The class introduces logic to: - Detect if /etc/resolv.conf is managed by netconfig or NetworkManager. - Verify user customizations using MD5 checksums (compatible with netconfig) and header analysis. - Preserve custom configurations by moving them to a static file and creating a symlink to prevent future overwrites by network tools. - Support resolver setup for both the target root and live systems. - wicked2nm: explicitly remove sysconfig-netconfig (bsc#1258112) If sysconfig-netconfig is not removed, when dropping wicked, NetworkManager still use `netconfig`. But `netconfig` doesn't find its `/etc/sysconfig/network/scripts/netconfig.function` file, which lead to an error like: ``` NetworkManager[926]: /usr/sbin/netconfig: line 574: log: command not found ``` - Fixed log message for already mounted system - Use download-in-advance mode for zypper migration This reduces the risk of download problems while inside of the upgrade process. This Fixes jira#PCT-1170 - Add explicit SLES15 migration target check Instead of falling back to the default with a misleading warning message, make sure to check for this target explicitly beforehand - Bump version: 2.1.33 → 2.1.34 - Init tools/migrate with debug set to false - Fix Zypper class command output handling The way zypper is called always redirects all output, stdout and stderr into the main log file. Because of that the variables self.output and self.error are always empty. This commit fixes the command call in a way that stdout and stderr are multiplexed such that the caller data can be captured by the python call and the data gets appended to the log file too. This commit also drops the unused and due to the redirection always empty self.error variable. For multiplexing the tee command gets used which impacts the returncode of the actual call. In order to get the correct exit code we use set -o pipefail. - Fix validation of zypper migration result Up to now the assumption was that any situation in which zypper migration cannot migrate the system returns with an error, meaning exit code != 0. However, this assumption is wrong. There are condition in which zypper migration only indicates a problem with a message saying 'No migration available' and the call return with a successful error code = 0. This causes big trouble for the DMS in a way that it continues running its services which all assumes the migration to the next major release was performed. It misleads readers of the log file into the wrong direction and the worst it causes modifications to the host system when it was not migrated. This commit makes sure that the migration stops and treats the above message as an error condition. - Fix quoting - Add time info to backup directory So far only the date information was part of the backup directory. However, if multiple migration attempts happens on the same day, this would overwrite the data. - fix: ensure `logger.handlers` is empty in tests Loggers will only be initialized once; thus if the logger test is scheduled after other tests, loggers might already be initialized and the logger creation code path might not be taken, causing the logger test to fail. - XFSv4 precheck - Fix migrate tool for container based migration The tool overwrites an eventually existing /etc/sle-migration-service.yml config file. This commit fixes it by using yq and inplace updates - update SLES16 DMS image to 16.1 - Fix md device detection Fix detection of device mapper layered rootfs. Software raid disks are not found by findmnt if the actual rootfs is one or more layers down. This commit fixes the detection if we need to pass along the rd.auto cmdline option to let the boot process activate them. - update SLES16 DMS image to 16.1

  • Release Date
    Jul 18 2026
  • References
    Bugzilla: 1260340, 1263889, 1258112
  • Type
    recommended
  • Severity
    important

cloud_download Downloads

SUSE Linux Enterprise Server 16.0 ppc64le
SUSE Linux Enterprise Server 16.0 x86_64
SUSE Linux Enterprise Server 16.0 aarch64
SUSE Linux Enterprise Server 16.0 s390x
SUSE Linux Enterprise Server for SAP applications 16.0 ppc64le
SUSE Linux Enterprise Server for SAP applications 16.0 x86_64