The minutes for 22nd January were approved.
Adding some virtual hardware, particularly processors, seems to have improved things a bit. Further investigation will follow.
The GPU and cluster computing page lists the rooms which have GPU machines. It didn't give 9.02 accsss rules, but does now. Documentation on how to use these machines is still pending.
This led off on a tangent to some apparent lack of awareness of computing.help, even though it's mentioned on the login screens. Could these be made any clearer? Carol is thinking about putting up some posters to advertise these things.
US are making progress persuading people that old machines can be removed.
B.03 is pretty full at the moment. CCL are pending.
Actions added or revived:
Re-defer to February 2021, by which time many things will be clearer and testing will likely be necessary for various reasons.
There are just a few reboots for kernel upgrades pending. These are being chased in a separate action.
There's a project looking at logging more generally.
Report from Computing Executive Group
Reports from Units
Reminder: if you want to benefit from automatic shutdowns you do have to configure your machines appropriately.
The new EdLAN-DB replacement is expected to go live on 24th March. This will affect the way Support set up managed Windows machines. It is a point&click thing, and will be tailored by IS. Instructions are expected...
Alison will circulate a reminder to the CSOs
There was general agreement that we could do with improving documentation
rfe and its use.
This provoked quite a bit of discussion. Regardless of the merits of the particular use case here, it's unlikely to be unique. We are at least aware of the problem now.
Do not leave this to the last minute!
It would probably make sense now to raise the loghost's DynaFileCacheSize.
We could also now excise any DICE-level SL7.5-isms.
What about the migration of our remaining users off staffmail over to O365?
Do we need to make any changes as a result of this?
How will our inbound mail be affected? We currently make our primary MX inaccessible from outside to force remote MTAs to go through one of the lower-priority ones for spam filtering.
What about smtp.staffmail? What do we currently tell our users? What do we have in computing.help?
Services should speak to Graeme and/or Kenny to find out more about the implications for us.
Support are doing a thorough tidy-up of the inventory. Everyone will be receiving an email asking them to verify that they still have the machines listed against them. This will be used to drive "sustainable computing" decisions.
The policy regarding the age of kit is active and is retrospective. The policy regarding the number of machines is active but is not retrospective (but machines caught by the "age" policy won't be replaced if this would conflict with the "number" policy).
Topics for discussion
Chris suggested holding over his Trello demo for a meeting where the previous business had been less.
The next meeting will be on Wednesday 26th February 2020, in AT-7.07, at 10:00.
Please contact us with any comments or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh