[Halld-offline] Outage Alert: move of /work/halld and /work/halld2, Wed. Nov. 29, 9 am to 10 am

[Mark Ito]

Folks,

We are scheduled to make the move of our work space from a Lustre-based 
system to a ZFS file system. This will happen tomorrow (Wednesday, 
November 29) starting at 9 am. The move should take about an hour.

N. B.: *during this time, you will not be able to write to the Hall D 
work disks*. Reading may be intermittent as well.

Some notes on the new configuration:

Both /work/halld and /work/halld2 will be on the same partition. The old 
/work/halld2 will actually appear as a new directory, 
/work/halld/halld2. A merge of the two systems is not a practical 
option. There will be a soft link, /work/halld2 pointing to the new 
/work/halld/halld2 directory, so legacy scripts should work.

Since there will be no distinction in the underlying hardware for the 
old /work/halld and /work/halld2, there is no more need to distinguish 
how they should be used. Small files and SQLite files are OK anywhere on 
the new work disk.

To reduce future confusion, you may want to move your files out of 
/work/halld/halld2, but that is not required at present.

Any scripts that use

   /lustre/expphy/work/halld or /v/lustre2/expphy/work/halld

instead of

   /work/halld

or that use

   /w/halld-scifs1a

instead of

   /work/halld2

will break. In other words, the server-specific directory names will 
change (as they must).

Let me know if there are any questions or objections to this schedule.

   - Mark

-- 
Mark Ito, marki at jlab.org, (757)269-5295

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20171128/244e5a18/attachment.html>