![]() |
Apologies for absence
Stephen and Chris had sent their apologies.
minutes of the last meeting.
George pointed out that the order for the AT rack referred to in the minutes had been placed in the purchasing queue but not yet ordered.
Report from Computing Executive Group
This agenda item was not taken.
Reports from units.
Infrastructure.
The new LDAP master server, franklin, was brought into service on Thursday May 17th, and all appears to have gone smoothly. Toby has been keeping a close eye on the LDAP logs since the changeover. Toby has commented that he believes that the new server is faster. Toby is still investigating the strange LDAP replication problems mentioned last time. He will also be investigating how to tune LDAP so that it is more responsive and resilient.
A new CA (Certification Authority) root certificate for Informatics was generated just over a fortnight ago. All managed DICE machines should have automatically regenerated their copies. In connection with this there are a couple of things that will need to be looked at eventually; code signing (RT 27718 ) and trust anchor (Bugzilla 708).
Network kit for Appleton Tower Levels 6, 7 and 8 has arrived and will shortly be configured.
We will be releasing rooms 1206b, 1206c, 3316 and 1501 in JCMB (all of which are labs) for use by others in the College this Friday, May 25th. George will check that the relevant switches are either removed or reconfigured.
Alastair mentioned that we will need to address the state of the server room in Appleton Tower basement.
Managed Platform.
Alastair reported that the checking of desktop hardware against the FC6 release is almost complete and that they have started checking server hardware (where we have spare kit of a particular type). The use of fibre channel with FC6 has been checked already.
The divine component will not be supported under FC6 (although it will still be available for George and others to use). The support for configuring laptops during a suspend will also be dropped.
George asked about a warning message from AFS during the booting of FC6 systems. AFS complains that the syscall table in the Linux kernel is unwritable; Alistair will look into this.
The unit will start to install new RPM repository slaves since the current slaves share the same hardware as the console servers and the two services have different requirements with respect to the configuration of the disk.
The unit have also been reviewing their documentation; it is currently spread across too many separate locations. They intend to place all the LCFG level documentation on either the LCFG web server of its wiki and all the DICE level documentation will be rationalised on the DICE wiki.
Research and Teaching.
The unit have ported the vast majority of the 120 or so that are needed for teaching and research to FC6. Most of the approximately 20 packages still to do are java related; the base java and eclipse packages are however already done. The most problematic packages have been those associated with haskell. The unit members will shortly start to sign the packages that are built (this will simplify the recovery from any possible future compromise of machines). Alastair asked that packages also be built for a 64-bit environment wherever possible.
Condor is now deployed on all Appleton Tower lab.machines except those in the quiet lab. There have been problems with running condor on hyper-threaded cpus. Condor treats the machine as if it had two separate cpus and only associates the keyboard activity with one of them, leading it to treat the other cpu as usable for its own jobs.
The unit have upgraded the remaining FC3 servers to FC5.
Iain has been looking at the GPFS (General Parallel File System) kernel module; it doesn't work on FC5 but does work on RHEL (Red Hat Enterprise Linux). Tim thinks that it would be possible to get the module to compile under FC5 but doubts whether it would then work. There are also problems with the Xilinx USB kernel module on FC5 (the unit got it to compile but it didn't work). Alastair suggested that they try an earlier FC5 kernel and failing that a RHEL kernel.
Tim has vastly speeded up the genrep database report system by using rsync to copy the contents of the repository of conduits and include files to the database server instead of using ssh and cvs.
Tim has supplied Anna and Stuart with a list of processes that will be affected by the introduction of EUCLID.
Services.
Craig reported that the mail server crashed at 04:45 on Monday morning with a kernel error. Iain noticed the problem from home and brought the server back up remotely by 09:00. The unit will liaise with the Managed Platform Unit over diagnosing the problem.
The colour printer on level 5 of the Appleton Tower hasn't been in service for a couple of weeks. It had been physically damaged by someone using too much force when opening one of the covers. It is out of warranty and is not on maintenance so a replacement printer will be ordered.
User Support.
Since the last Operational Meeting the User Support Unit had handled 115 new RT tickets (equivalent to about 15 per working day) and resolved 61% of them. There had been a total of 124 tickets (including both new and existing tickets) resolved over the same period.
Ken reported some figures showing resolution rates against age of RT ticket:
Ticket Age D | Percentage Resolved (or rejected) |
---|---|
1 week < D < 2 weeks | 80% |
2 weeks < D < 1 month | 85% |
1 month < D < 3 months | 91% |
3 months < D < 6 months | 96% |
6 months < D < 1 year | 96% |
1 year < D | 99% |
In just over a week several hundred old computer accounts, belonging to non-Informatics students who have not been enrolled on an Informatics course for over a year, will be deleted.
No FC5 desktops need to be rebooted to pick up the new kernel. The unit's seven remaining servers that needed to pick up the new kernel have now all been rebooted.
AOCB
George kindly agreed to chair the next Operational Meeting on Wednesday June 13th in Ken's absence (assuming George is available).
Please contact us with any
comments or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh |
![]() |