News: systems
Datacentre electrical maintenance: Fri 13 Jan 2023
29 December 2022
Work to upgrade power resilience in the Computer Lab's datacentre (GN09) will take place from 10th until 13th January 2023, requiring a shutdown of some electrical circuits in our datacentre for most of 13th January. This will require powering off a number of (mainly) research servers, and some network equipment connecting a larger set of research servers. Owners of affected servers have been contacted. We have some (very limited) capacity to provide temporary power to some machines if arranged in advance. More …
Filer maintenance: Tues 28 Sep (rescheduled)
20 September 2021
Update: this work has been rescheduled for the afternoon of Tuesday 28th September due to contractor unavailability. More …
ssh users: upgrade of ely / slogin-serv system
21 July 2021
The TSS 'ely', the main machine in the slogin-serv.cl.cam.ac.uk ssh service, is being upgraded as there is a security vulnerability in its kernel. If things don't work, please try slogin-new in place of ely. More …
www-dyn / www-dyn1 rebuilt on Ubuntu 20.04 - www-dyn2 (Fedora 9) shutdown
27 March 2021
The www-dyn service, provided by www-dyn1 runing Fedora 9, ceased working a year ago (the www-dyn2 and www-dyn3 services continued to run). More …
Datacentre cooling works
26 May 2020
An upgrade to the cooling capacity of the department's datacentre (GN09) will take place starting on 1st June 2020, for two weeks. More …
svr-ssh-0 now Ubuntu 20.04 - please try it if you use slogin-serv
3 May 2020
slogin-serv.cl.cam.ac.uk is provided by three servers, mostly by "ely". One of the two alternatives, svr-ssh-0.cl.cam.ac.uk, has been upgraded to Ubuntu 20.04. If you routinely use slogin-serv, please try using svr-ssh-0 and email sys-admin if tere any problems. Thanks. More …
Guidance on using Zoom
24 April 2020
UIS has now published guidance on how to use Zoom securely for online videoconferencing. More …
Launch of Microsoft Teams
19 March 2020
Microsoft Teams is now live. It is the recommended and supported online collaboration tool for the University. Teams is integrated with other Microsoft 365 (formerly known as Office 365) products. More …
Mailing list for announcing filer migration disruption
5 September 2019
The migration of filespaces from elmer to our new NetApp servers is likely to start sometime after 24th September and will take several weeks to complete. It will be necessary to disable access to individual parts of the filestore while they are being copied. For example, your home directory, bigdisc space and any group filespaces you use are very likely to be disrupted on different days. More …
filer alias updated
4 September 2019
In a previous announcement, I said that the Windows file server name \\filer would change from being an alias of \\elmer to a dedicated server that would offer redirections to the appropriate destination for each item in the filesystem. This will allow filespaces to be migrated to the new NetApp fileserver a few at a time without end users having to know which is where. More …
Changes to the www-dyn{,1,2,3} service
30 August 2019
The facility used by the dynamic user web page service to securely access the filer relies on a feature of the old filer which is not available on the new ones. More …
Old autofs maps to be removed
29 August 2019
At the beginning of August we introduced a new set of autofs maps, and as far as we can tell they are working well. It is possible that a few machines are still using the old maps, but this will inevitably lead to problems because they are no longer being maintained. We therefore intend to remove the old maps from the LDAP server on Tuesday 3rd September. This change will ensure that all machines should follow the changes that will be necessary to migrate filesystems to our new filer. More …
New NetApp filer
22 August 2019
The department has been running a NetApp filer called elmer as its primary filestore since 2002. Since the original installation, there have been two major hardware replacements, and on each of those occasions the transition was relatively seamless. We had about a day of downtime, during which the filesystem was transferred over and an almost identical service launched on the new hardware. It is now time for a third upgrade, but this one will be a bit different. The purpose of this article is to explain what we are planning. More …
New autofs maps
1 August 2019
In preparation for migrating to new filer hardware, we need to make changes to the automounter mappings used on managed Linux machines to access NFS file servers. These mappings are used to implement paths mostly beginning with /auto, and most frequently accessed through symbolic links such as /homes. /anfs and /usr/groups. The maps are made available to client machines from our LDAP server. More …
Chiller failure
16 May 2019
A pump in the temporary chiller failed at around 3AM. The current
suspision is that a pump tripped / failed, possibly due to a power
glitch.
GN09 warmed up to about 50C.
Remedial action taken at 7AM which brought the temperature down to
around 40C.
The pump was identified and reset at around 8:15, and temperature
returned to normal. More …
New Active Directory domain DC.CL.CAM.AC.UK
4 February 2019
The department is migrating from an old Active Directory domain to a new one. If you are still using somewhere the domain/realm/NetBIOS names of the old one, please check the current documentation and change to the new one accordingly: More …
Old print server being retired
7 November 2018
The old print server (version 1-6-2) will be retired in stages over the next week or so. Most users will not be affected, as they are already using our new server. The only people who need to take action are those who used Bonjour to load printers via the old server, which will be mostly Mac users. These printers will cease to work and they should be deleted and replaced with printers using the new server. At the moment a Bonjour user will typically see two entries for each printer, e.g. More …
Network outage on June 2nd
4 June 2018
At 09:12 on Saturday morning, a new switch in our data centre failed. In doing so it bridged together all of its connections to the rest of the network. Since switches have multiple uplinks for resilience, this created a loop in the network. More …
New Ubuntu 18.04 LTS slogin-new test server
28 April 2018
A new Ubuntu 18.04 LTS server is now available, available via the service name slogin-new and slogin. To ensure that the Ubuntu 16.04 LTS server is used, connect to slogin. Comments to sys-admin please. More …
Start of change of default domain on Linux systems
9 March 2018
The department is moving from the 'AD' domain to the 'DC' domain. Linux systems use the DNS to decide which the default should be if none is given. We shall be moving certyain test hosts over to DC in the vulnereable period on Tues 20th Feb, and depending how things go, at later vulnereable or non-vulnerable periods. More …
Server Room Cooling Failure
28 February 2018
Overnight between the 27th and 28th of February the building cooling failed. This was due to the pipes on the roof freezing due to very low temperatures. Emergency actions was taken at in the middle of the night to shut down enough machines to prevent a total outage of core services. The cooler has now been inspected and declared beyond repair. This is because it had no anti-freeze in the circuit. More …
New Remote Desktop Service
26 February 2018
The previous remote access Windows machines ts01 and ts02 have now been retired. In their place is a newer version running on Windows Server 2016. The local unix command cl-rdesktop has been updated to point to this service. To login you must authenticate against the DC domain either by quoting your identifier as DC\crsid or as just crsid. Windows or MacOS users should connect to clrds.cl.cam.ac.uk to access the new servers. More …
Problem with new VPN certificate corrected
22 February 2018
Some people using the VPN service at vpn.cl.cam.ac.uk may have been experiencing difficulties since the server certficate was updated on 2018-02-13. We issued a warning applicable to Windows users, but we forgot to point out that some Linux and Android clients require a local copy of the server certificate to verify. More …
SE18 due to be cleared by 1 June 2018
15 February 2018
The deadline for moving equipment out of SE18 has been set to the 1st June 2018. The removal of the air-conditioning has been arranged to start them. More …
ely reboot in vulnerable period 28th Feb
15 February 2018
This reboot is necessary to complete the installation of some security updates. Apologies for missing the reboot on the earlier advertised date. More …
CL VPN certificate replacement
6 February 2018
The SSL certificate used to authenticate the CL-specific instance of the UIS VPN service at vpn.cl.cam.ac.uk is scheduled for routine replacement on Tuesday 13th February at about 08:00. The purpose of this certificate is to give the VPN client the assurance that it is communicating securely with the correct service. More …
Filer software update Tuesday 9th Jan 2018
4 January 2018
The next scheduled vulnerable period (2018-01-09 0700-1000) will be used to perform a software update on the NetApp filers elmer and eldo. Elmer is the main filer directly visible to users and eldo provides the backing store for virtual machines. More …
UPS move 10th January 2017
1 January 2017
On the 10th January the UPS was moved. This left core services vulnerable to building power outage until it is recommissioned later that day. More …
Microsoft Imagine Username Changes
10 October 2016
The usernames on the Microsoft Imagine site (formerly known as Dreamspark and before than as MSDNAA) have been changed to be of the form [Javascript required] for all users. The effect of this change will be that you receive notice of your new account details and the old one will no longer be available. If this causes any difficulties for you please contact [Javascript required]. More …
Website upgrade Tuesday 4th October
30 September 2016
On the morning of Tuesday, 4th October 2016 we will upgrade the web server software behind the main Lab web site. More …
Dropbox
15 September 2016
The UIS have recently made available reduced cost access to Business Dropbox available via normal purchasing means. More …
Terminal Server access now requires VPN
29 August 2016
Following a number of brute force password attacks against our Windows Terminal Server hosts these have now been restricted to only accept connections from clients within the University network. If you need to access them from outside that network then first connect using a VPN. More …
ssh-dss .dsa user ssh host keys to slogin-serv
18 July 2016
The new Ubuntu 16.04 slogin-serv services do not
accept the legacy ssh-dss (DSA) public key algorithm. Any users
having problems with their user ssh keys should try connecting to
slogin-old using the -v switch to check
which key is used. If it is .dsa, please generate a new
key and add it to .ssh/authorized_keys, check that it
works on slogin-serv, then remove the old .dsa key and
its .ssh/authorized_keys entry. Remember never to copy
the private key - it should only ever remain where it is generated. More …
slogin server partial upgrade to Ubuntu 16.04
7 July 2016
Subsequent to the move of slogin-serv to Ubuntu 14.04, another server has been added, which runs Ubuntu 16.04, accessible as slogin-new.cl.cam.ac.uk, which has been in service for a while now without any reported problems. It is now being added to the slogin-serv.cl.cam.ac.uk pool, so users may get Ubuntu 14.04 or 16.04. The hope is that it should not matter, and when we are confident that 16.04 is suitable, the pool will move to being just 16.04. If you find something which requires 14.04, please connect to slogin-old.cl.cam.ac.uk and email the details to sys-admin More …
new slogin-new - consider using slogin-serv
22 May 2016
The service name 'slogin-new.cl.cam.ac.uk' is used for
the newest slogin pilot or service for the department (see Time Sharing Syetsms). A new system is in
development, so anyone wanting to use the previous 'new' server
sandy should use
slogin-serv.cl.cam.ac.uk (using ssh user keys) or
sandy (using Kereberos). TGT server users should check
their configuration files. More …
TexLive 2015 now available on the filer
25 June 2015
The latest version of TeXLive is now available on the filer in /usr/groups/texlive. Using the standard TeXLive naming convention, it is in the directory 2015. More …
Disruption to Xen clients
28 May 2015
Our Xen system was disrupted last night. It is believed to be mostly fixed, but users should check their systems. A rogue process caused problems with the Lab routing, which caused the filer to be inaccessible, causing some domUs to deem their file systems broken. It is thought that most effected systems have been restarted, but there may still be problems which need fixing. Please email sys-admin if any help is needed. More …
DHCP servers upgraded
9 March 2015
The main Lab DHCP servers have been replaced. This should not have changed things for clients, so if there are any problems, please report them to sys-admin. More …
Another power glitch 2015-03-03 03:06:24
3 March 2015
This time it killed a switch ad (at least) two discs. More …
West Cambridge power problem 2015/02/23
24 February 2015
There was a power problem on the West Cambridge site effecting some of our systems in the WGB and some in the UIS at around 9PM on Mon 23 Feb. Please report any anomalies to sys-admin. More …
slogin server upgrade
21 February 2015
As there have been no reports of problems with the testing slogin server we plan to swap the DNS pointer over the weekend spanning the end of February and the start of March at which point the old instances will be available for a short time as slogin-old. More …
NFS version 4
11 November 2014
Support for NFS version 4 has now been enabled on the main NetApp filer "elmer". The change is upwards compatible in the sense that the older protocol remains available as well. However client systems which support multiple versions will normally use the highest version available from a server, so many machines will see a change. More …
WGB power cut 2014/11/09 17:43
9 November 2014
There was a power cut in the WGB just before 17:43 for about ten seconds. See UIS report for their timeline. More …
internal 'sysdata' web service fist to move to NFS sec=sys
27 October 2014
http://sysdata.cl.cam.ac.uk/ is the first of the Lab web services to move to Ubuntu servers running Apache 2.4 using NFS sec=krb5 to access the web pages from the filer. Please contact sys-admin if there are any problems. Other services are likely to move over in the near future. More …
Matlab updated to R2014b
23 October 2014
The version of Matlab used on lab Linux systems has been updated from R2013b to R2014b. More …
Removal of avahi-daemon from Ubuntu 14.04 systems
22 October 2014
It appears that the avahi-daemon causes the gtk+ library on Ubuntu 14.04 systems used by many GUIs (e.g. firefox and chrome) to hammer the print spooler by opening hundreds of simultaneous connections, causing the GUI to crash. To avoid this, we are removing the avahi-daemon on all Ubuntu 14.04 systems. If you manage a machine which needs it, simply add 'avahi-daemon' to /etc/user-config/bundles as usual. and it will be auto-installed. More …
Internal Lab facilities may not be available
24 September 2014
Certain commands used to do “privileged” things within the Lab may not be working for a while. If some feature you particularly need does not work, please email details to sys-admin More …
condor pool moved to NFS sec=krb5
16 September 2014
As emailed to all recent users of condor, the execute machines have now moved to NFS sec-krb5. The documentation has been updated. If anyone has problems, please email sys-admin. More …
network and Xen problems
5 September 2014
A misconfiguration in the networking kit caused an outage yesterday afternoon. This caused many Linux Xen VMs to loose access to their root filesystems, so they mounted them read-only. After rebooting some of them, one of the Xen Servers lost the ability to reboot VMs, so itself had to be rebooted, causing all its VMs to crash. Many of the filesystems were deemed "unfixable" by fsck, so fsck had to be manually run before the systems rebooted. Please let sys-admin know of any remaining problems. More …
new slogin-otpw server - generate a new pad
8 August 2014
mgk25 has just released version 1.5 of OTPW
which does not need to use $HOME for the pad. This allows per-host
local pads, removing the problems on NFS sec=krb5 systems. The DNS name
slogin-otpw now points at the testing slogin server. Any users
need to generate a new pad by logging in and running
cl-otpw-gen. More …
Ubuntu 10.04 lucid libc bug
6 August 2014
A libc bug involving nscd may cause some Ubuntu 10.04 systems to fail in various ways. The quick fix is to run "cl-asuser service nscd stop" and "cl-update-system" to install a fixed libc. Ubuntu 10.04 is getting a bit old now, and 14.04 LTS is stable - it is straightforward to install a fresh Ubuntu 14.04 onto a Lab system while it is in use, and the reboot into the new system - email sys-admin if you would like your machine upgraded. More …
Wireless outage
18 July 2014
The report I have received from the UIS about today's wireless outage is slightly inconclusive. It seems that there was a malfunction of the central controller that looks after our access points, causing them to reboot and go into a default unconfigured state. They were initially rescued by moving them to another controller, but have now been put back as they should be without further problems. More …
Testing new slogin / ssh-remote server
2 July 2014
The main Lab Linux TSS is currently running CentOS 5. We plan to move to Ubuntu Server 14.04 LTS, so have set up a testing server "sandy" (AKA slogin-new and ssh-remote-2.cl.cam.ac.uk). Please try it out and report any problems to sys-admin. More …
Xen Server upgrade to 6.2
1 May 2014
The last main pool of Virtual Machines will be upgraded (6.0 to 6.2) soon, probably over the Bank Holiday weekend. Attempts to do a "smooth" migration failed for the last three pools, so this time it will not be attempted. The downtime of each machine should be short, but there will be some unavailable disruption to logged in users. Please email sys-admin if you have any particular availability requirements over the weekend. More …
www-dyn(1) upgrade
28 April 2014
The server providing the www-dyn1 and www-dyn services is to be upgraded. Users are recommended to check that any pages they publish work on the www-dyn2 and www-dyn3 servers. See http://www.cl.cam.ac.uk/local/web/personal.html#www-dyn for full details. More …
Further NTP related changes
24 February 2014
With the new routers, Lab machines lost the ability to NTP probe machines on the internet, as we lost stateful UDP reflexive rules. With the recent NTP amplification attacks, we have added rules which mean that machines on the internet cannot NTP probe most Lab machines (our NTP servers should still be accessible). Peerings (where source and destination ports are both 123) should still work. More …
Purge of unused VLAN 100 IP addresses
18 January 2014
As part of a campaign to clear the use of the original lab subnet, IP address on VLAN 100 in the subnet 128.232.0.0/20 which are believed not to have not been used for a year have been removed from the DNS. Further pruning is likely in future. If you believe an address has been retired but is still needed, please email sys-admin More …
Wordpress upgraded to 3.8
23 December 2013
The instance of wordpress used for http://www.cl.cam.ac.uk/news/ has been upgraded to 3.8. Please report any problems to sys-admin. More …
DNSSEC for cl.cam.ac.uk
15 November 2013
On Monday 18th November 2013 we plan to add a DS RR for cl.cam.ac.uk to the cam.ac.uk zone. This will mean that resolvers using DNSSEC will check our information, and ignore it if the information does not have a valid signature. Apart from making our information more secure, it also has the ability to take cl.cam.ac.uk out of the DNS if something goes wrong. Please report any DNS problems with cl.cam.ac.uk More …
Filer running slowly
18 September 2013
The filer is running slowly at the moment. This is because we're copying the entire filesystem to the replacement machine. The ultimate goal is to make everything faster, but in the short term we have to suffer some pain. More …
Purge of legacy automount information
17 April 2013
It is believed that the last systems using the old "indirect" autofs LDAP tables have been shut down or moved to use "direct" filer mounts, so the information is no longer included in the LDAP data. If anyone observes failure of the automount system (rather than missing krb5 tickets) please report the machine and the directory to sys-admin. More …
Xen Server problems - please report anomalies
24 July 2012
We have had a number of problems with one of our Xen pools. If any VMs appear to be behaving oddly, please contact sys-admin. A dom0 (actual machine which hosts VMs) had problems, so was taken out of service and replaced with a hot-spare. One VM (only!) appears to have problems running on the dom0, but there may be others lurking. More …
CentOS5 (Linux server) network service problems.
24 July 2012
A Redhat package problem caused CentOS 5 machines (Linux Servers) to loose access to certain network services. One of our two mail servers may have rejected email between 04:15 and 04:30, when the problem was spotted and the failing MTA was taken out of service. More …
DNS resolvers now validating
26 May 2012
The CL DNS resolvers have DNSSEC validation enabled. Please report any DNS problems to sys-admin. More …
Remote backups resumed
8 February 2012
The regular remote backups that were suspended on 30th January have now resumed. In particular home directories are now up to date on the remote server and should continue to be backed up hourly. More …
IPv6 address changes
8 February 2012
There has recently been a change to the IPv6 address allocations used in the University, though at present this affects relatively few people. The reason for the change is that the original allocation to Cambridge was not thought to be large enough; a larger allocation was subsequently obtained but this entails phasing out the old addresses. More …
Problem with remote backups
1 February 2012
Since the morning of Monday 30th January, we have had a problem with the offsite backup system; a number of filesystems, including the one containing user home directories, are not being copied to the offsite server. More …
Network problem 2011-11-25
28 November 2011
There was a major problem with the CL internal network during the afternoon of Friday 25th November, believed to have been caused by a faulty switch. There were a number of consequential side effects on other systems, which we now believe have been dealt with. More …
Unix group cleanup
25 October 2011
On departmental Linux systems, many Unix groups have been added locally, to facilitate groups of users jointly editing files. Use the shell command "id" to see which groups you are a member of. To avoid malfunctions, we now need to clean up older groups with gid value below 500. This requires your help. There are two sources of groups: More …
CentOS 5.7 released - kernel lock bug fixed
15 September 2011
Although not yet announced on the CentOS web site, CentOS 5.7 has been released, which includes a new kernel which fixes the bug which caused the machine to OOPS if an NFS file lock was attempted. As such, there will be a campaign of rebooting CentOS 5 machines, such as slogin-serv and miscellaneous other Lab servers. As CentOS systems automatically do "dot" upgrades, systems should automatically install the new kernel, and all that should be needed to move to the new kernel is a reboot. More …
Wireless network replacement - phase 2 complete
29 July 2011
Phase 2 of the wireless network replacement is now complete, adding a number of mid-corridor access points. More …
Phone outage Thursday 28th July
28 July 2011
The Computing Service send their apologies for the telephone system outage this morning. A small and fairly routine reconfiguration was made to the West Cambridge routers during this morning's CUDN vulnerable period, ironically with the ultimate goal of improving network resilience. Unfortunately a maintenance script failed to run correctly and the voice network became disconnected. Once reported (by mobile phone), the communication channels worked well and the fault was attended to promptly. More …
Wireless network replacement - phase 1 complete
20 July 2011
The initial phase of the replacement wireless network installation is now complete. All of the old access points have been decommissioned, and 24 Aruba access points have been installed. More …
Password requirements change
19 July 2011
Passwords created or changed from now onwards will need to be of higher complexity than was permitted previously. Details of the requirements can be found at http://www.cl.cam.ac.uk/local/sys/access/kerberos/. Existing passwords will continue to work as normal - you only need to be aware of the extra restrictions when you choose to change your password. More …
Wireless network replacement
13 July 2011
On Monday and Tuesday of next week (18th/19th July) the long awaited upgrade of the wireless network in the building will take place. There will be disruption to wireless access as old access points are replaced with new, but I hope that there will be connectivity available from most places at most times. More …
New package signing key
8 June 2011
A new key (3294FB03) has been generated to sign all Lab packages (RPMs and .debs). This has been installed on most Lab managed machines. Private machines which use the Lab repositories will need to register the new key. Details available on the WiKi More …
World IPv6 day - 2011/06/08 - possible network service problems
7 June 2011
On world IPv6 day there is a possibility of disruption of network services due to machine with dual stack not working correctly. Although we will be making small changes, most of the changes are outwith our control. If you experience problems, please contact sys-admin so that we can monitor and try to sort things. When we last enabled local services via IPv6 as well as IPv4 there were problems due to some Windows machines thinking that they have IPv6 connectivity, trying to use the IPv6 address, and failing. We believe that the particular problem was fixed, but there may be others. More …
svn services changes
17 April 2011
As mentioned in centos-5-machines-upgrading-to-5-6 CentOS machines have upgraded. They now run svn 1.6.11, so the svn-1-6-6-serv service is being withdrawn. More …
CentOS 5 machines upgrading to 5.6
10 April 2011
CentOS 5 machines (mostly servers rather than user workstations) are automatically upgrading to 5.6. It should be mostly transparent, but the mailer (exim) has a security fix, so the old config file causes email to fail, including email saying that there is a problem. Some machines may have problems with the upgrade, and Virtual Machines may install a kernel which fails with NFS locking. If you use a CentOS machine (see /etc/issue) and it has problems, please email sys-admin. More …
WordPress upgraded
1 April 2011
The old WordPress software used for these pages has been upgraded. The theme has changed, but other things should be much the same. More …
Network outages and consequences
17 March 2011
As of about the 11th of March, we have been having a number of problems with servers crashing, machines misbehaving, network dropouts, things going slowly, etc. We believe this to be due to some sort of local networking problem with the switches. Latest update: The supervisor in use on the central switch was was swapped Tue Mar 22 09:52 and the problems did not recur within four hours. More …
Unix groups now also available in AD
21 February 2011
All Unix groups are now also maintained as a set of Security Groups in the Active Directory. The membership is controlled from the Unix groups and additions and deletions to maintain the groups in synchronisation are performed hourly during the working day. More …
slogin-serv move to NFS sec=krb5
7 February 2011
Any users of the "slogin-serv" service should be aware that changes
are to be made which may require them to run "kinit" to
access their home directory. Lab Linux machines are moving
from using NFS with sec=sys (which trusts the client OS user
information to auth users) to sec=krb5 (which requires the user to
have a kerberos ticket to indicate what access they should have) to
improve the security of user files on the central NFS server. Users of
new or upgraded Linux systems will be used to sec=krb5, but it's a bit
different on servers. Some hints and tips are available on the
NFS
sec=krb5 web page. We will be moving the main "slogin-serv"
service over starting the week beginning 2011/02/07. Initially the
name 'slogin-serv' will point at the existing sec=krb5 instance. Users
may need to type "kinit" (or equivalent) to get access to the filer.
Users experiencing problems should login to 'slogin-old' (or
'slogin-serv9' or 'ramsey') and then email sys-admin explaining the
problem. When we are confident that things are basically working, one
of the old servers ('ramsey') will move over to sec=krb5. When we are
happy that everything is working, the temporary name 'slogin-old' will
be removed, and 'slogin-serv9' will be retired. Users are encouraged
to try the new service in advance by connecting to ssh-remote-0, and
send any feedback to sys-admin More …
Web server changes
10 January 2011
Two small changes have been made to the Lab web server that may break some existing web pages, in particular those using other character encodings than UTF-8. More …
Network flood
2 December 2010
A traffic flood from a research group machine caused some network performance problems between sometime in the small hours of Thursday 2nd December and about 09:40 when the source was identified and disconnected. Our use of VLANs to separate traffic did a reasonable job of confining the problem, but there were a few knock-on effects; in particular one of the DNS resolvers found itself behind an overloaded link causing slow name lookups. Some overnight jobs may have misbehaved as a consequence. More …
"wgb" wireless network
19 November 2010
The "wgb" wireless network is running very slowly at the moment and is congested. The reason for this is that the ADSL connectivity that lies behind it has been capped at 128kbit/s because of excessive usage under Demon's fair usage policy. We are allowed 200GB in any rolling 30-day period, and it seems that this has recently been exceeded. More …
Network switch failure
21 October 2010
A network switch providing connections in the north east corner of the first floor failed shortly after 04:00 this morning (2010-10-21). It is one of two switches in the wiring closet so some machines and telephones in the area would have lost all connectivity whilst others were working normally. More …
Mail server replacement
15 October 2010
The last exim3 MTA (whittlesey) is about to be replaced by an exim4 one (sutton). This should not cause any disruption, but please email sys-admin if you suspect any problems. If email is completely failing, do phone or come to see us. Shutting down the last exim 3 server allows us to drop legacy features, and when we are convinced we won't need to roll back to either of the old exim3 servers, we'll be in a position to make use of new facilities in exim4. More …
slogin-serv* hosts IP addresses changing
11 October 2010
The IP addresses associated with the slogin-serv* servers is changing as they are moving to a subnet for machines using NFS sec=sys. This may cause ssh to warn that that the IP address differs, but that the host key matches. This is not a problem, so type "yes" to connect. e.g. More …
WGB power outages
11 August 2010
The WGB lost power on Wed 11th Aug 2010 and three times on Sat 14th due to the main 240v breaker in GN17 tripping, probably caused by a problem with the temperature sensor, from reports of other sites. It is likely to require a six hour power down at some future time to fix it (unrelated to the per-corridor shutdown on 16th August for monitoring work). More …
Caller ID presentation on Lab phones
6 August 2010
The new University phone system has the ability to present "caller id" on external phone calls as the actual DDI number of the phone, rather than the generic university switchboard number. This makes it easier for people to call you back. Whether the number actually reaches the called party is dependent on a number of technical and administrative factors; at present it will work for most calls to UK landlines. Unfortunately it won't work for most calls to mobile numbers. More …
Demise of Computing Service usenet server and email to it
29 July 2010
The Computing Service have announced that the University usenet news server will shut down at the beginning of next month (August 2010). Their announcement suggests alternative usenet sources. More …
Wireless network migration plans
19 July 2010
This is a status report on the planned upgrade to the wireless network, and a request for testing and feedback on the migration strategy. More …
Eduroam
14 June 2010
I have received a couple of samples of the new wireless access points that we intend to install over the summer. I have placed one of them so that it should provide some Eduroam coverage to the Street area. More …
New package signing key
8 June 2010
A new key has been generated to sign all Lab packages (RPMs and
.debs). This has been installed on most Lab managed machines. Private
machines which use the Lab repositories will need to register the new
key. On a debian based system use "apt-key
add" and on an RPM based system use "rpm
--import", using the key on the lab web server, e.g. one of:
wget -O -
http://www.cl.cam.ac.uk/deb/cl.cam.ac.uk-archive-keyring.gpg | sudo
apt-key add -; wget -O - www/deb/KEY|sudo apt-key add -; sudo rpm
--import http://www.cl.cam.ac.uk/RPM/RPM-GPG-KEY-CL; More …
Filer problems May bank holiday weekend
1 June 2010
Between sometime in the evening of Sunday 30th May and early morning on Tuesday 1st June the performance of the NetApp filer "elmer" was so badly degraded as to be effectively unusable. The problems affected various services, including the main lab web server. More …
Migrating the filer to a new IP address - second stage
31 March 2010
On 22nd February I wrote about the start of a campaign to the migrate the NetApp filer "elmer" to a new IP address. There have been a few hiccups, but by and large the migration has gone smoothly. The vast majority of machines migrated to the new address of their own accord, and over the last few days the stragglers have been cajoled into moving by more or less drastic means. More …
Migrating the filer to a new IP address
22 February 2010
We will shortly be making a change to the DNS to change the default IP address that is used to access the main filer, elmer. The motivation for the change is to deal with some security issues that are currently preventing the use of the latest version of the NFS protocol. More …
Web browser proxy configurations
4 February 2010
The Computing Service have recently announced that their web proxy service is to close. We have not recommended use of this service for many years, but there is still some legacy from the time when it was strongly recommended. Part of that legacy is a collection of proxy configuration files in our WWW server that some older browsers may be configured to consult. More …
Web server logs
7 December 2009
The web server logs (subsets of which are accessible using the cl-personal-www-log command) are now being recorded in "combined" log format rather than "common". This format adds two extra fields, giving referrer and user-agent information. More …
Changing the name on telephones
26 October 2009
Until now, changing the name displayed on our VoIP telephones has been a big deal, involving an interaction with the telecomms office. We now have access to an experimental facility allowing us to change these strings ourselves. More …
cron job migration from legacy servers
21 October 2009
In preparation for decommissioning our older omnipotent servers, various cron jobs are being rehomed. This should not disrupt thing, but please let sys-admin know if there appear to be problems. More …
Linux kernel vulnerability tweaks
14 August 2009
Due to a Linux kernel vulnerability a number of machines have had mmap_min_addr set to 4096 or had modules.alias tweaked to disable some protocols. If any users suspect that this may have broken things on their system, please email [Javascript required] More …
Main Lab web services use a pool
7 July 2009
There are now four web servers running the same configuration which provide a pool from which the services (www.cl.cam.ac.uk, www.tex.ac.uk, etc) can be served. More …
Brief problem affecting certain home directories
12 May 2009
There was a period of about 20 minutes starting at about 09:30 this morning when certain users (those on "homes-1") may have had problems accessing their home directories. This was traced to a "garbage-in garbage-out" problem in the script that does the routine filer housekeeping, and corrected itself next time the script ran. The cause of the original "garbage-in" is still not known, but steps have been taken to detect it in future and abort the script before any damage is done. More …
Windows7 RC1 now available locally
11 May 2009
You can now download Windows 7 from the local MSDNAA site, www.cl.cam.ac.uk/MSDN, in both 32 and 64 bit versions. Please remember that since this is a release candidate it will stop working after a set period. It is only suitable for experimental use. More …
dns[01].cl auth NS to refuse "other" requests
7 May 2009
The authorative nameservers for cl.cam.ac.uk are moving to return REFUSED for requests for other domains rather than sending back AUTHORITY and ADDITIONAL RRs to help the client to find the requested RR. More …
new page on sec=krb5
5 May 2009
The new page nfs-sec-krb5 has some initial simple (!) examples of using NFS sec=krb5 - feedback welcome. More …
Tool to test ssh authorized_keys files
1 May 2009
Users can check that their ~/.ssh/authorized_keys files is
"acceptable" with
/usr/groups/admin/cmds/check-ssh-auth.pl. More …
Terminal Server changes
1 May 2009
We are about to release a new terminal server (ts01.ad.cl.cam.ac.uk) to replace the ones (waterloo & charingcross) that have been in service for the last 6 years. This runs on Windows 2008 rather than 2003 and so roaming profiles are not shared between the old and new services. In addition the version of Office on the new server is Office 2007 which has a radically different user interface to previous versions. More …
Web server upgraded
27 April 2009
The main lab web server was upgraded to Apache 2.2 on 27 April 2009. Please check any particular configurations (e.g. via .htaccess) that you rely on and contact webmaster@cl if you spot any problems. More …