Skip to end of metadata
Go to start of metadata

Previous Outages & Events

Planned Activities

2017-10-19 06:00

2017-10-19 10:00Monthly routine maintenance (extended)NCSARoutine patching and reboots, firewall firmware updates, server firmware updates.All NCSA-hosted resources except for Nebula.


2017-10-19 06:00

2017-10-19 14:00qserv-master replacementNCSA

qserve-master will be down so that systems engineering can finish configuring the new server and xfering files. Status updates here:IHS-378 - PDAC qserv-master upgrade IN PROGRESS.

qserv-master will be down for this entire period


2017-07-20 04:00

2017-07-20 08:00

Monthly lsst-dev maintenance

  • Routine system updates.
  • Note that this event will begin 2 hrs. earlier than normal
  • Note that the July 20 maintenance event will include permanently unmounting the remaining NFS filesystems.
  • IHS-365

SeeIHS-365 - scheduled maintenance for NCSA-hosted dev machines (July 20, 2017, 4:00 - 8:00am Pacific) DONEfor details

verify-worker31 suffered a failure and will be out of commission for a while
2017-06-22 (06:00)2017-06-22 (10:00)

Critical Kernel upgrades


Upgrade kernel and system packages to address Stack Guard Page vulnerability. See also:IHS-324 - Emergency update of all NCSA hosted dev machines DONE

All NCSA hosted resources (except Nebula).

UPDATE: 08:00 PT - Outage is being extended till 10:00 PT.

Outage was completed at 10PT. Some nodes didn't come back. See ticket for details.


2017-06-05DAQ installationNCSA

Mike Huffer postponed to a later date.

Pushed back - yet to be rescheduled

2017-05-18 (06:00)

2017-05-18 (08:00)

LSST monthly maintenance

  • Kernel upgrades and reboots in the LSST dev environment
  • Permanently unmount the old NFS home filesystem. This completes the decommissioning of the NFS home filesystem.
  • Install of Unbound local caching resolver software as recommended by NCSA Security
  • Remount of remaining NFS exports in read-only mode. Users should migrate any old data off of NFS in preparation for the final NFS decommissioning. NFS is expected to be turned off entirely on July 20, 2017.
  • Upgrade to latest MySQL 5.5.56 on


2017-05-04 09:302017-05-04 10:00


lsst-dev file systems full



lsst-dev filesystems / and /home filled up at approximately 09:30. This was a result of inode usage from another process

The admins freed up inodes to make the filesystem responsive again.

Admins are currently tracking down the root cause.

2017-04-27 13:112017-04-27 14:20


Nebula outage

glusterfs crashed due to this bug, so no instances could access their filesystemsAll instances running on NebulaNeeded to reboot the node that systems were mounting from, but took the opportunity to upgrade all gluster clients on other systems while waiting for a reboot. Version 3.10.1 fixes the bug. All instances with errors in their logs were restarted.

2017-04-20 (04:30)

2017-04-20 (09:30)

LSST monthly maintenance


This event is cancelled so as not to interfere with Early Integration Activity #03 being held at NCSA April 19 & 20.

nothing bad happened

2017-04-17 (13:41)

2017-04-17 (13:53)


lsst-dev login node down


Users unable to log in to lsst-dev.

Probable cause is that the root file system filled up due to excessive logging

2017-03-27 (22:00)2017-03-29 (14:00)Blue Waters maintenanceNCSA

Due to maintenance of cooling infrastructure at NPCF, Blue Waters will down during this period. Cray will also take this maintenance window to perform some system updates at the same time.

Systems that will be down

  • Slurm cluster compute nodes will be powered down for the duration of the outage.

Systems that will remain up

Qserv nodes ( lsst-qserv-* ), SUI nodes ( lsst-sui-* ), Bastion node ( lsst-bastion01 ) should remain online during the outage.  

However, if temperatures in the NPCF rise too high, we will be forced to shut these down as well. I've been told that this is a low-probability scenario and we will be given time to do graceful shutdowns. In the unlikely event that this happens, it will be communicated through the DM Slack channel and also posted here.

All systems normal

2017-03-23 (0800)

2017-03-23 (1300)

NCSA Nebula OutageNCSANebula will take an outage to balance and build a more stable setup for the file system. This will require a pause of all instances, and Horizon being unavailable.

Nebula is back to normal.

2017-03-16 (0430)2017-03-16 (0930)LSST monthly maintenanceNCSAGPFS filesystems will go offline for entire duration of outages. Some systems may be rebooted, especially those that mount one or more of the GPFS filesystems.

Aug. 24, 06:00

Aug. 24, 13:30

LSST Dev infrastructure upgrades

  • Infrastructure upgrades to all LSST-dev resources:

    • lsst-dev01
    • lsst-db
    • lsst-xfer
    • lsst-web (lsst7)
    • lsst-dts
    • lsst-dbdev environment
    • lsst-daq test stand
    • lsst-dbb (data backbone) test stand
    • lsst-elast (elastic compute) test stand

Completed successfully

Aug. 24, 06:00

Aug. 24, 07:30

LSST Dev patching

  • Regular maintenance (Puppet releases and patching) of NPCF-based systems
  • IHS-416

Completed. Note the separate status message for "LSST Dev infrastructure upgrades", which includes system in NCSA 3003 and is scheduled for 06:00 - 15:00.

Maintenance on the following:

  • Cluster Services
  • Verification cluster
  • Prototype DAC
  • effected systems include: adm01, backup01, bastion01, monitor01, object*, qserv*, sui*, verify-worker*, test0*

Aug. 22, 06:00

decommissioning of lsst-dbdev machines

  • The lsst-dbdev nodes are old, obsolete and unused. They'll be powered down, but kept for a few months in case a need for the data on them arises.


2017-06-15 06002017-06-15 0730

Deploy unbound LSST cluster nodes (verify-worker*, qserv*, sui*, bastion01, test*, backup01)

NCSADNS resolving may have a short (~30 mins) delay.Updates deployed successfully via new puppet module. All tests passed.
2017-02-22 14152017-02-22 (1615)Nebula Gluster IssuesNCSAAll Nebula instances paused while gluster repairedNebula is available.
  • No labels