We have identified 703 home directories on home.eecs.berkeley.edu that are not under immediate, active use, and will be migrating them to the new file storage beginning late tonight, January 22, 2016. Active use was assessed using timestamps of top level home directory, last Active Directory login, last CIFS access, NFS mounts and netstat output. We anticipate that most of the migrations in this first batch will take just a few minutes, and that the whole first batch will be complete before Monday.
You may view the list of home directories that are part of this first-batch migration here: List of home directories to migrate
During the migration, access to the old location will be disabled; you will be able to access the new location immediately, as files are copied over.
Many users may never notice a difference. If you manually mount your home directory, please read on for changes. If you have questions about some of the technical details, please contact your system administrator(s), or contact help@eecs.
1) NFS exports will change. AMD/automaps will be updated accordingly. Note that this implies a re-mount.
usernames Old New
a-h home:/vol/home/eecs/<user> home10:/home10/<user>
i-q home:/vol/home/eecs/<user> home20:/home20/<user>
r-z home:/vol/home/eecs/<user> home30:/home30/<user>
2) CIFS/Windows access will change similarly. Active Directory will be updated accordingly, so that corresponding H: drives will be mounted from the new location.
usernames Old New
a-h \\home\<user> \\home10\<user>
i-q \\home\<user> \\home20\<user>
r-z \\home\<user> \\home30\<user>
We plan to schedule the migration of the remaining home directories on home.eecs in the coming week. Please stay tuned for more information.
UPDATE
[2016-01-22 21:42:17 | Lars Rohrbach]
A small change to the list of affected directories reduces the count to 678. The list available by the link above has been updated.
UPDATE
[2016-01-25 19:45:34 | Lars Rohrbach]
We have completed the migration of this first batch of home directories.
During this procedure, we unintentionally caused some unexpected service interruptions for some users:
* Twenty-one accounts were migrated which were not on the original list of “first-batch” accounts. While this occurred because their home directories were determined to not be in active use, it was not our intention to surprise anyone with an unannounced migration and the temporary unavailability that goes with that.
* Automount map updates for targeted users were pushed out all at once, even for users who would have been skipped because their home directories were found to be in active use. The automount map update caused their home directories to become immediately unavailable. This was corrected late Friday night, and subsequent automount map updates were staged for users with successful migrations.
A list of accounts successfully migrated so far is available here:
Migrated Accounts after First Batch
Resolved as of 2016-01-23 23:00:00