sPHENIX discussion of software

[pinkenburg <pinkenburg AT bnl.gov>]

    Hi folks,
    
    rcf will be down for maintenance on Tuesday Dec 1st. The details are
    below but the main effect will be an interruption of email service
    of rcf accounts. For those of us who use rcf as email provider (me
    included) don't expect to get any emails or be able to send any
    emails during that outage.
    
    Mattermost, BNLBox and our external (and internal) websites will
    also be down.

The cssh campus gateways (the ones we use to tunnel into BNL for our time cards or the intranet.bnl.gov) will stay up.

In preparation, the batch queues will be closed on Nov 30, 18 hours before the outage. So please do not submit 2-3 days jobs after Nov 29th. Initially it was discussed to stop accepting jobs 3 days in advance to guarantee that all jobs could finish but this was seen as not necessary since most jobs run shorter than that and we would loose 2 days of computing.

afs will be down also for remote users. cvmfs is cached on the OSG stratum 1's, so likely it will stay available (as long as you don't access a file they need to fetch from the stratum 0 at BNL).

Chris


-------- Forwarded Message -------- Subject: [Rhic-rcf-l] SDCC Downtime 12/1 Date: Fri, 20 Nov 2020 16:22:15 -0500 From: RACF Computing Facility Staff <announce AT rcf.rhic.bnl.gov> To: rhic-rcf-l AT lists.bnl.gov, sdcc_users-l AT lists.bnl.gov, bnl-shared-tier3-l AT lists.bnl.gov

The SDCC's internal network will be brought down for important maintenance on December 1st, from 6:00 AM to 6:00 PM EST. This will impact the accessibility/usability of ALL SDCC services including:
-Connectivity to SSH gateways (note the campus gateway, cssh.rhic.bnl.gov, will stay online), NX servers, and interactive nodes (existing connections will be closed)
-General services including RCF mail service, Mattermost, Discourse and RT
-Batch processing
-Disk storage access, including remote AFS access, access via our SFTP gateways, NFS, GPFS, Lustre, and BNLBox
-Tape/HPSS storage access
-Access to the SDCC Kerberos service
-SDCC web server access

In preparation for this maintenance, the HTCondor pools on the SDCC HTC compute farm (RCAS/RCRS, SPOOL, and SPAR hosts) will be closed to new jobs starting at noon EST on 11/30. Any user batch or interactive processes still running at 6:00 AM on 12/1 will be terminated. Our HPC clusters (IC, KNL, Skylake, ML, and CFN gen4) will not start Slurm jobs which will run past 6:00 AM EST on 12/1.

Once the maintenance is completed, mail service will resume and a notification will be sent to all users.

--
This message has been forwarded from the RACF announcements page.
Recent messages are available at:
https://www.racf.bnl.gov/Facility/RACFNews/announce.html
_______________________________________________
Rhic-rcf-l mailing list
Rhic-rcf-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/rhic-rcf-l