Archive:Skaro

From Tardis
Revision as of 23:48, 26 October 2011 by Amazoph (talk | contribs) (Argh! Someone didn't update the pages after 'renaming' the links)
Jump to: navigation, search
Note: Retired Systems pages are generally outdated and for historical interest only. Unless otherwise stated assume the system/hardware does not exist.
The hardware listed is still in the rack destined to become our new router. See: Admin Projects

Skaro is named after the home planet of the Daleks and, at times, the centre of the Dalek Empire.

Configuration

Skaro is a Sun Netra T1 AC200 donated by Charles Leahy (charles). It's a 1U UltraSparc IIe box with 768MB of RAM and two 10K Seagate 36GB Cheetahs. It is also one of the few machines in our rack which is installed with the correct rackmounting kit.

Hosting

Skaro has been designated as the new mail server and could also be used as a sparc shell server.




Details on lifetime as router:

Ace
Operating System: Debian Linux (Lenny 2.6.26-2-sparc64)
Hostname: ace.tardis.ed.ac.uk
IPv4: 193.62.81.1 (int), 129.215.160.171 (ext).
IPv6: 2002:81d7:a0ab:1::1 (int), 2002:81d7:a0ab::1 (ext).
Hardware:
Netra T1
Primary Services:
Router
Other Information:
IPv6 Tunnel

Ace has been commissioned (as of 30/09/10) as Tardis' router. At some point it may be put in a HA setup with King. RIP sungem.

Hardware

Ace (formerly mara) is a Sun Netra T1 AC200 donated by Charles Leahy (charles). It's a 1U UltraSparc IIe box with 1GB of RAM and two 10K Seagate 73GB Cheetahs. It is also one of the few machines in our rack which is installed with the correct rackmounting kit.

OS & Storage

It is running Debian Lenny 2.6.26-2-sparc64 and is set up with two RAID1 partitions: one 3GB root and another ~70GB partition for /usr.

Routing

Still todo :)

Things that are broken

There have been occasions where ace's networking has died completely, and it had to be rebooted. On closer inspection, this message was seen in syslog:

12:47:21 ace kernel: [955103.491804] eth0: RX MAC fifo overflow smac[02010400].

It's likely that large amounts of data or connections (such as when a system decides to do a full apt-get update/upgrade) are causing this to occur. Hopefully someone can get a kernel memory dump the next time this happens, so we have some more detail for a bug report.