FAS Research Computing - Subset of compute nodes in row 7c not running at full speed – Incident details

Cannon Compute Cluster (Holyoke) experiencing degraded performance

Status page for the Harvard FAS Research Computing cluster and other resources.

Cluster Utilization (VPN and FASRC login required): Cannon | FASSE


Please scroll down to see details on any Incidents or maintenance notices.
Monthly maintenance occurs on the first Monday of the month (except holidays).

GETTING HELP
https://docs.rc.fas.harvard.edu | https://portal.rc.fas.harvard.edu | Email: rchelp@rc.fas.harvard.edu


The colors shown in the bars below were chosen to increase visibility for color-blind visitors.
For higher contrast, switch to light mode at the bottom of this page if the background is dark and colors are muted.

Subset of compute nodes in row 7c not running at full speed

Identified
Degraded performance
Started 2 days ago

Affected

Cannon Cluster

Degraded performance from 4:15 PM to 12:00 AM

Cannon Compute Cluster (Holyoke)

Degraded performance from 4:15 PM to 12:00 AM

Updates
  • Identified
    Identified

    Approximately 100 compute nodes in row 7c (holy7cxxxxx) came back up at the wrong clock speed. This is a known issue with some older nodes that can occur after power dip or loss. These nodes will need to be physically reset in the chassis to clear this issue.

    Jobs are running on these nodes so they will need to be drained of jobs before we can do this. As such we are marking the cluster partially degraded.

    However, please note that this is a fraction of the total nodes and should not impact the average user.

    This affects some specific owned partitions:

    doshi-velez - 2 of 2 nodes
    shakhnovich 2 of 3 nodes
    blackhole/blackhole_priority - some % nodes
    hernquist - some % nodes
    pehlevan - some % nodes
    seas_compute - some % nodes
    tambe - some % nodes
    shared - some % nodes