Header Search Widget
IRIS

Instructional & Research Information Systems

Lars Rohrbach

NetApp filer upgrade on December 1

by

As part of our preparation for migration to new NetApp file storage hardware, we will be upgrading the OS on our existing Network Appliance file servers at 5pm Tuesday, December 1.

The OS will be upgraded from Data ONTAP version 8.1.2 to 8.2.4. Upgrade will be performed for project.eecs and home.eecs first, and then for cluster.eecs and inst-fs.eecs.

No general service interruption is expected, as our servers are configured in High-Availability pairs. Windows/SMB/CIFS users may experience a hiccup, if they have an active file transfer at the time of the upgrade; such a file transfer would terminate and need to be restarted.

Work is expected to be complete by 6:30pm.
[Read more…] about NetApp filer upgrade on December 1

Nov 2015 Windows Server Patching

by

Microsoft will release their patches for the month on Tuesday, November 10th, 2015.

PRINT, WINTERM and WDS will automatically patch and reboot Wednesday, November 11th, at 3:00 am.

WINTERM may be unavailable till 10:00 am Wednesday, November 11th, depending on additional software updates.

Resolved as of 2015-11-11 08:00:00

Campus DNS change caused some NFS mount trouble

by

On Wednesday, October 28, campus migrated to a new set of Domain Name System (DNS) Infoblox appliances. In the process, domain names under their control were changed from the previous standard of mixed-case (e.g. violet.Berkeley.EDU) to all-lowercase (e.g. violet.berkeley.edu). This change was unannounced.

DNS names are “case insensitive”, but some applications that use DNS names are aware of case. In particular, NFS exports of our filesystems from home.eecs, project.eecs, and cluster.eecs, which are managed by netgroups in NIS, rely on case-sensitive name resolution to allow or deny access. The campus change caused exports to machines in subdomains such as Millennium.Berkeley.EDU, CITRIS.Berkeley.EDU, and IST.Berkeley.EDU to suddenly be denied. We have updated our own NIS netgroup entries to reflect the case changes, so expected NFS access has resumed.

Other applications that may be affected by this case change include web and database applications.

Campus has indicated that they have no plans to roll back the changes in Infoblox. They recommend that users update their applications to the new standard.

At this time, our independently-managed subdomains (EECS.Berkeley.EDU, CS.Berkeley.EDU, Banatao.Berkeley.EDU) continue to be mixed-case.

Installation of New NetApp File Storage

by

We are pleased to announce that we are moving to the next generation
of file storage in EECS.

Our current Network Appliance file storage hardware is approaching
its end of life, and our storage needs continue to grow.
Maintaining our own file storage systems has been valuable to research
and departmental operations, keeping costs down while providing centralized
storage with backups, snapshots, CIFS/NFS access, and room to grow.

Over the coming months, we will be migrating to a new cluster of NetApp
file servers, which will include disk-based backups and mirroring of current
data to the San Diego Supercomputer Center. Some changes to mounting and
some service downtime are expected, but details are still being worked out.
Look for upcoming announcements on how this migration might affect you
and your access to home.eecs, project.eecs and cluster.eecs.

Brief service interruption for gateway.eecs, cricket.eecs

by

At 4:30pm today, we will briefly shutdown gateway.eecs and cricket.eecs, to move them to another rack in the same room.

Downtime should be only 5-10 minutes.

gateway.eecs is used for some outgoing SMTP connections, such as might be configured in Thunderbird or Outlook.

cricket.eecs is used in our network database to display port usage information.
[Read more…] about Brief service interruption for gateway.eecs, cricket.eecs

Microsoft patches affecting some domain clients

by

We have received multiple reports of people having trouble logging into EECS-domain Windows machines this morning, following application of the most recent Microsoft patches. Clients receive a message indicating: “The security database on the server does not have a computer account for this workstation trust relationship.”

Re-joining the EECS domain fixes this, but a simpler solution is also reported to work:
* Unplug your ethernet connection (wireless users may have a physical switch to disable wireless).
* Login with your EECS account as normal (it will used cached credentials).
* Reconnect ethernet cable (or enable wireless).
* Logout, and then test that you can successfully login normally (while connected) again.
[Read more…] about Microsoft patches affecting some domain clients

Network disruption following the Soda Power Outage

by

Following the completion of yesterday’s Soda Power Outage, and some network maintenance that was also completed during the outage, we discovered there was a network issue affecting a number of Cory Hall clients and wireless access points.

Some hosts affected include:

  • cusg.eecs.berkeley.edu
  • www-bsac.eecs.berkeley.edu
  • venus.eecs.berkeley.edu
  • cory.eecs.berkeley.edu
  • imail.eecs.berkeley.edu
  • terra.eecs.berkeley.edu

We will update here as our investigation continues.
[Read more…] about Network disruption following the Soda Power Outage

Campus network issue affecting IRIS services

by

Campus IST is investigating reports of network packet loss. This network issue affects IRIS services such as the IRIS website, and our network database and roster, since they depend upon IST databases and servers hosted on IST virtual machines.

For up to date information on this outage, please see http://ucbsystems.org/
[Read more…] about Campus network issue affecting IRIS services