[Halla_running] Counting house disk wipe

[Ole Hansen]

Dear colleagues,

TL;DR: /chafs2/work1 will be wiped soon. Start backing up your data by 
March 31 or lose them.

In preparation for SBS running, we are planning to upgrade the disk 
systems in the counting house. The first disk area to be upgraded will 
be /chafs2/work1. This disk is used primarily for storing online 
analysis results (ROOT files etc.)

Unfortunately, it will not be possible to preserve the data on 
/chafs2/work1 when upgrading. To avoid losing your data, therefore, it 
needs to be backed up elsewhere.

Here is the current space usage on /chafs2/work1:

46T    apar
19G    APEX
555M    dvcs
498M    prex
287G    triton

A cursory inspection shows that most of these directories have not seen 
any activity for at least a year, often much longer. I will presume 
these data are abandoned unless I hear otherwise.

The apar directory, however, in addition to occupying most of the space, 
holds quite a number of files that were created as recently as last 
month. Therefore, would the PREX/CREX collaborators please have a look 
and decide if anything needs to be kept.

To back up files from /chafs2/work1, there are several choices:

- Directly back up to MSS from the counting house machines, using the 
usual Jasmine tools (jput etc.) When doing so from the group accounts 
like "a-onl" or "apar", please temporarily copy (or generate) the 
scientific computing certificate of a user who has MSS write permission 
to the respective group account's home directory.

- Copy the data to /volatile/halla/<your-experiment>/<some-subdirectory> 
and take it from there. The /volatile disk is visible from all counting 
house machines. Both the "a-onl" and "apar" accounts are members of the 
"a-parity" Unix group, which should help with write access permission. 
File ownership can be changed later.

- Physically attach a portable USB disk to one of the counting house 
machines and copy the data you need. You'll need assistance from someone 
with root access, e.g. Bob Michaels or myself.

We would like to proceed with the upgrade in the next few weeks. Please 
reply by *March 31* if you want any data kept. For large-scale backups, 
like the 46TB of parity data, you should be prepared to supervise the 
backup jobs yourself.

Best regards,
Ole Hansen


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halla_running/attachments/20210323/24f8cbf4/attachment.html>