There was a brief internet outage today (April 1st) around 11:20am, and it lasted for about 10 minutes. There was an issue with our border router – it was unresponsive possibly due to an attack. After a reboot, the router functioned normally again. Further investigation will entail.

For those of you that are using our EECS spam service, here is an update.

After months of service degrade, the vendor finally concluded that our spam appliances have been “poisoned” inadvertently due to the overall setup within EECS. That is why we are seeing less than optimal success rate, indeterministic tagging of emails, etc.

The only way to resolve this is to do a complete reset of our spam configs. We will do this on Tuesday, 3/29/05, at 9am. There will be no downtimes. We will do as much pre-training as possible, especially for false negatives. The extra taggings will continue.

As a user, you might see some false positives/negatives, which you should send them to helpdesk (help@eecs), along with the complete header. If you opted out of the auto-filing on IMAP server, this will not have any impact. Please also let us know if you encounter any other problems.

—-

To All Windows Domain Users:

As you may know the departments remote installation server for Windows, ris.eecs.berkeley.edu, has been down for several months. Starting Wednesday, March 23rd, we will resume serving an XP SP2 image to all authenticated domain users.

Unfortunately, Microsoft is still at work on a fix for our Windows 2000 image. Assuming we get this fix, we will resume serving the Windows 2000 image. Until we get this fix, you can continue to use the install CD’s available at the departments help desk to install Windows 2000.

Thanks for your patience during this down time. If you have any questions about using the XP SP2 image, please see the department’s Help Desk in 395 Cory Hall or send mail to help@eecs.

The departmental NIS, DNS and time server, rhea.cs, will be down for routine maintenance beginning Monday, March 21st at 5:00 PM. The maintenance is scheduled to end at 5:30 PM, but we expect rhea to be back on-line by 5:15 PM.

We expect that rhea (and cronus) will not need to be off-line on Wednesday, March 23rd when repairs to the cooling tower system in Cory Hall will effect the 165 Cory machine room.

Please make your plans accordingly.

On Tuesday, March 22, the EECS wireless network will be down to load a maintenance release of code. The outage window will be from 06:00-08:00. AirBears service within Cory, Soda and Euclid will also be down during this time. Please note that only portions of the net will be down at any one time. However, users may experience more than one service interruption during this window. We regret any inconvenience this may cause.
[Read more…] about Upcoming Wireless Outage

PP-CS is calling for non-critical computer systems in room 165 to shutdown on Wednesday, March 23, for repairs to the cooling tower system in Cory Hall. We ask you to volunteer to shut down your computer equipment for the hours of 8am – 4pm. We believe we can keep the room cool (if we have enough volunteers) for those who can not shut down their computer equipment or servers. It is expected that the ambient temperature will rise during this repair from 70 deg. F. to 80 deg. F. in room 165.

Also, PP-CS will be testing some standby cooling this Thursday, March 17, in room 165 and will require no action on your part.

Please advise if you will be able to help us by shutting down your equipment. If we can not shutdown any equipment we will likely reschedule a full shutdown later this year.

Please note this is not related to the industrial water shutdown on Tuesday, March 22. This industrial water shutdown should not affect room 165 cooling system.

Thank You,

Scott McNally
[Read more…] about Cory Cooling Curtailment

In order to comply with the recent campus IT security policies (http://itpolicy.berkeley.edu), some changes will be made to the main EECS outgoing email relays. Please customize your settings accordingly. If you have further questions, please contact help@eecs.

1. If you use “gateway.eecs.berkeley.edu” as your outgoing email relay, these changes will not impact you. If you are not, we strongly suggest that you update the outgoing email relay to “gateway.eecs.berkeley.edu”. Please note that most ISPs have started blocking port 25, so if you have problem sending email via “gateway.eecs.berkeley.edu”, the helpdesk can assist you.
2. Effective 3/29, if your outgoing email relay is “smtp.eecs.berkeley.edu”, authenticated relay will start. If your client (outlook, netscape, thunderbird, etc.) uses ‘secure connections’ “when possible”, you might see a warning about mismatch of SSL certificate (which you can ignore). If you use Eudora, this mismatch cannot be ignored, and you will need to change your outgoing relay and might need to re-import the Entrust certificate chain. Again, the helpdesk can assist you.
3. Effective 4/26, only authenticated email relays will be accepted by our outgoing email relay servers. If your client (outlook, netscape, etc.) does not have the ‘secure connection’ set, please set it to “when possible” to allow authentication.
4. If any servers within EECS that use “smtp.eecs.berkeley.edu” as a relay server, this will also cease to function after 4/26. Please have your system administrator make the appropriate changes. More information will be sent in the near future.

The Entrust certificate is attached.

Attachment:  entrustchaincert.crt

On Wednesday, March 16, from approximately 7pm-10pm, IDSG will be upgrading the master LDAP server, ldapmaster.eecs.berkeley.edu. During that time, attempts to write to LDAP will fail. None of the read-only slaves will be offline, and as such, most LDAP-dependent services should continue to function.

The following should be the list of things that won’t work during the upgrade.

• LDAP password changes (either through the helpdesk password reset webpage, or the IMAP webmail client on imap.eecs)
• Roster account creation/modification/deletion
• Ports database changes

NIS password changes (ssh’ing to passwd.eecs) and Windows AD password changes will continue to work.

The RIS server continues to be unavailable, although IDSG has high expectations that it will be back some time this month.
Microsoft is close to solving our problem, however additional testing is needed. Install disks continue to be available at the help desk, thank you all for your patience.

The Instructional UNIX file server suffered a hardware failure around
10am Saturday March 5th. Please see the Instructional web site for the latest updates.
[Read more…] about Instructional UNIX home dirs and email are inaccessible