Whole Root Zone Migration Fails (6608767)

Whole root zone migration, including branded zone migration fails. Zone migration and copy, with contents, for a whole root zone might result in the following errors:

• Whole root zones are large in size. Tarring, compressing, and splitting whole root zones take a lot of time. Solaris Container Manager has a small timeout period after which it tries to reattach the original zone.

• Zones are split in fixed sizes of 2Mbytes. The size of the whole root zone is in Gbytes, so the sub- archives are large in number. The awk command does not support string lengths exceeding specified limits.

• The older zone archives remain on the server. During download, some unwanted files might be downloaded as a part of the zone image.

• The target system disk space should be at least three times the size of the zone. Inadequate disk space results in failure while extracting the archives.

In addition to these errors, the following issues might also occur:

• The zone contents are migrated completely, but might not be in the required state to attach to the target. This issue might occur even when Solaris Container Manager is not used for migration of the zone. In such a case, the zone may remain in the installed state, but will not boot.

• The web server installed with Sun MC, might fail. In such a case the whole file transfer will also fail.

Workaround: Ensure that the following conditions are complete:

• es-config -l root -M scm-container the command and profile updates have run successfully

• The hosts have the same hardware configuration and the same the release of OS installed on them

• There is sufficient disk space on the hosts.

• The /var/opt/SUNwsymon/downloads/ directory does not contain any files. Delete the files, if any.

• Wait at least for 2 minutes after clicking on Finish in case of zone copy or migration. Sometimes a blank window might remain on the screen. This window is harmless and can be closed forcefully.

• The /var/opt/SUNWsymon/cfg/scm-zonecopy.properties file has predefined values. The predefined migratePollInterval (600000) can remain as it is. The effective sleep time is equal to pollInterval x retry. Increase the default retry value from five to ten for a larger zone.