There will be a semi-urgent maintenance reboot of our mail server and our terminal server "plimpy" today, Friday, 2009-Dec-04, at 6pm. So if you are using one of our thin clients, please save your work and log out before 6pm. Expected downtime durations: approximately 15 minutes for the terminal server "plimpy" and approximately 30 minutes for the mail server.

Additionally there will be a restart of our LDAP authentication database at the same time. This can cause logins on managed workstations or connections to file servers to fail temporarily, but should take no more than a minute or so.

Update, 7pm: Maintenance over. All systems back to normal. Terminal server was back in time, mail server needed a file system check so the downtime took slightly longer than expected. LDAP restart was a matter of seconds.

On Monday, November 16th, starting at 6pm we will migrate the D-PHYS IMAP and POP3 server to new hardware. Since this also includes converting all 450 GB of mailboxes on that server, the migration will take several hours. Update Tuesday, 2:50pm: Everything works again. The migration took about twice of the expected time due to some complications.

During the migration it will not be possible to access any of the D-PHYS e-mail services (receiving, sending or accessing your mails, webmail). E-mails sent to D-PHYS during the migration will be held back on the sender's e-mail servers.

After the migration you may need to resynchronise your e-mail client (e.g. Apple Mail or Mozilla Thunderbird) with the new IMAP server. Please read our mail server migration wiki page in case you use one of the e-mail clients mentioned above, use your own procmail filter rules or experience other problems after the migration.
(more…)

On Thursday, 24th of September 2009 starting at 5pm there will be a maintenance downtime of the D-PHYS mail server. Estimated duration is one hour. During the downtime, the mail server will not be able to receive or send e-mails nor will it be possible to log in via IMAP, POP3 or SSH. Update, 6pm: Everything back to normal. (more…)

A security-update on one of our clusters failed unexpectedly and caused a downtime of our authentication server and printer server which needed manual intervention. All services should be back again by now. We apologize for the inconvenience caused by this. Update 18:00: The thin clients work again, too.

On Thursday, the 27th of August 2009 around 16:00 we will start upgrading the operating system and the webserver software of our main webserver zwoelfi. This may lead to a downtime of many websites at D-PHYS including www.phys.ethz.ch and isg.phys.ethz.ch. Update 21:20: All services should be back to normal again.
(more…)

There will be a short maintenance downtime of our authentication server this evening at 7pm. This means that you won't be able to log in to any managed D-PHYS workstation or D-PHYS service for about 5 minutes. Update 19:01: Maintenance successfully over.
(more…)

For a long time we used so called wildcard SSL certificates for most of our web sites and services. Especially newer web browsers issued warnings about them. We recently exchanged the wildcard certificate on https://www.phys.ethz.ch/ and other websites to a new type of SSL certificate listing all websites hosted on that machine. Since today also our infrastructure servers (e.g. https://registration.phys.ethz.ch/), have such a new SSL certificate.

If you have installed our CA certificate as described in our article How to configure web browsers, the new certificates won't cause any more security warnings. If you have not installed our CA certificate in your web browser, a warning about a new unknown certificate will pop up for many of our SSL-enabled websites every time we add a new website. Because a new website means that we have to generate a new SSL certificate for the whole web server the new site is hosted on. We therefore recommend to install our CA certificate in your web browser to avoid those annoying warnings on D-PHYS websites.

We had a partial DHCP outage this morning because the failover system between our two DHCP servers malfunctioned. The problem is currently solved only preliminarily by disabling the failover system and one of the two servers. To get both DHCP servers working again, we scheduled a DHCP server maintenance for today, Friday, 30th of January 2009, starting at 5pm. There may be short DHCP service outages during the maintenance. This means that initially connecting to the network as well as renewing the DHCP lease of your computer may fail for a few minutes. Update 21:45: The DHCP service is working again including the failover system.
(more…)

For kernel security updates and other maintenance work, there will be a downtime of the e-mail server of the ETH Department of Physics from 17:00 on for about one hour.

(more…)

On Tuesday, the 2nd of December 2008, beginning at 6pm, we will migrate the services on our current authentication server to our high-availability cluster. During the migration you will not be able to login on any of our managed workstations, home and group drives or web services. Existing login sessions on workstations should survive the downtime, but it is usually better if you log out before migration. Additionally, *.dhcp.phys.ethz.ch dynamic hostnames will not work either.

Update, 22:30: Migration done. Everything should be back to normal.

Update, Wed Dec 03, 09:30: We found and corrected a problem with the web browser proxy.pac.