sdcc_users-l AT lists.bnl.gov

Subject: Scientific Data & Computing Center

List archive

[[Sdcc_users-l] ] Update on the BNL electrical grid activity on 12/30

From: SDCC Announcements <announce AT rcf.rhic.bnl.gov>
To: rcfstaff AT bnl.gov, rhic-rcf-l AT lists.bnl.gov, sdcc_users-l AT lists.bnl.gov, usatlas-computing-l AT lists.bnl.gov, usatlas-ddm-l AT lists.bnl.gov, usatlas-users-l AT lists.bnl.gov, bnl-shared-tier3-l AT lists.bnl.gov
Subject: [[Sdcc_users-l] ] Update on the BNL electrical grid activity on 12/30
Date: Thu, 26 Dec 2024 18:13:02 -0500

Summary:
The SDCC was informed about a major electrical intervention on 12/30 on very
short-notice. Because this
intervention comes with a small risk of an electrical power loss, the SDCC
has decided to quiet-down its
services to minimize the risk of a long-lasting recovery process with limited
staff availability during the
BNL self-declared quiet period.

See updated times (highlighted in capital letters) in the text below.

Effective Time(s):
12/30/2024 9:00 am - 12/30/2024 6:00 pm

Group Responsible:
IT Fabric

Affected Area(s):
All SDCC services

Expected User Impact:
No access to SDCC resources (computing, storage and services)

Maintenance Type:
Planned Maintenance/Downtime

Description:
A critical maintenance/replacement procedure on the BNL main electrical grid
scheduled for Monday, Dec. 30th was announced
to the SDCC on very short-notice last week. This procedure is planned to
start around 12 noon and last approximately 4 hours.
We recognize this procedure is happening during the BNL-declared
&amp;amp;quot;quiet period&amp;amp;quot;, but a postponement
would incur increased costs to the Lab and potentially place this must-do
procedure during the start-up period for RHIC run 25,
which is deemed even less desirable than the current plan. BNL management has
decided to go ahead with the Dec. 30th
procedure, as planned.

This procedure requires transferring the power source from the electrical
utility to the back-up generator, with an UPS to
bridge the time gap (a few seconds) between utility and generator power, and
then remain on generator power for the duration
of this procedure. Because there is a small risk of failure during the
transfer process and in generator operations and because of
reduced staff availability during the BNL quiet period, the SDCC management
has decided to quiet down the facility resources
to minimize the chances of data corruption, service disruptions and hardware
failures, in the unlikely event that an unplanned
power outage occurs.

Quieting down means: 1) draining batch jobs (HTCondor and Slurm), holding new
ones from starting and stopping interactive
access to SDCC cpu resources on SUNDAY (DEC., 29TH) AT 3 PM ET and 2)
stopping all data read/write and movement activities
(disk and tape) on MONDAY (DEC. 30TH) AT 9AM ET.

Announcements to SDCC Liaisons and program/experimental PoCs will be made
when SDCC resources are fully available again.

SDCC Announcements page:
https://www.sdcc.bnl.gov/news-events/sdcc-announcements
Downloadable calendar invite of this event (.ics format):
https://www.sdcc.bnl.gov/announcements/make_ics.php?evt=1735254782

This item has been posted to RCF/USAtlas Staff, RHIC RCF, SDCC Users,
US-ATLAS Computing, US-ATLAS DDM, US-ATLAS Users, BNL Shared Tier 3 Users

--
This message has been forwarded from the SDCC announcements page.
Recent messages are available at:
https://www.sdcc.bnl.gov/news-events/sdcc-announcements
________________________________________________________________

[[Sdcc_users-l] ] Update on the BNL electrical grid activity on 12/30, SDCC Announcements, 12/26/2024