MPU Meeting Monday 31st October 2011
AFS automation project
Nothing happened.
LCFG Server Refactoring
Nothing happened.
Install Scripts
The report has been written, it will now be sent to COs for comments
and linked to from the Computing Systems blog.
Wake-On-Lan
Chris has been looking at the IS WOL client and server LCFG
components. He is also reading through the CGI script to see what it
does. We probably will need a separate cosign-authenticated CGI script
for Informatics users. Hopefully this web interface can use the same
configuration files which are generated by the WOL server component.
Simple KVM Service
The introduction of the service is waiting on a fix for the SL6 Fibre
Channel problems. The work on adding documentation is ongoing.
Server Upgrades
The first package cache server,
split, has been upgraded to SL6. The plan is to do
schiff this week. After that Stephen will look into upgrading the AFS build server,
lochranza, he will use the test server
prague to prepare the upgrade.
Miscellaneous Development
- hwmon
- Thanks to the addition of some new tools the LCFG hwmon component now supports H200 and SAS5iR RAID controllers. It was suggested at the last Operational Meeting that the hwmon code should be added to the relevant dice raid headers, possibly with a cpp guard-macro to allow an opt-out. Any warnings generated from running the script are now also passed back to nagios so we don't miss important issues.
- audit daemon
- Stephen has finished the LCFG components to manage the audit daemon and the audispd-remote plugin. This gives us the ability to log lots of information and send it to a central server in a secure manner. We need to put some thought into exactly what we want to monitor. We might want to deploy the basic logging to all servers. The components currently have no documentation, that needs to be fixed.
Operational
- Server compromise
- Stephen is still working on finishing off the report into the compromise of the SSH servers.
- LCFG sleep component
- There have been a few problems with suspend working but not reporting success, in this case the messages to rootmail will now only be sent if the
verbose
resource is set. There have also been a few problems with sleep failing after a shutdown has been scheduled. We should look at whether there is a way to spot that shutdown is running.
- LVM tools
- Alastair has added the LVM tools to the SL6 installroot.
- Visit to Durham Uni
- Stephen will be visiting Durham Uni Central IT service this Thursday and Friday.
This Week
- Alastair
- Complete KVM documentation
- Investigate multipath under SL6
-
Review spending plan (ServerBuying2011to14)
-
Add DNS problems with small servers onto MPU operational report
- Pass BIOS settings to USU
- Look at SL6 server stuff
- Chris
- WOL project - continue digesting James's CGI script
- Add hwmon to relevant DICE headers
- Review spending plan (ServerBuying2011to14)
- PD
- Stephen
- Review spending plan (ServerBuying2011to14)
- Complete report on ssh compromise
- Documentation for LCFG audit components and consider how to deploy
- Visiting Durham Uni
- Upgrading schiff
--
AlastairScobie - 31 Oct 2011
Topic revision: r5 - 08 Nov 2011 - 09:40:02 -
ChrisCooke