On Thu, 13 Jun 2013, Ganesh Sittampalam wrote:
> The real problem is more fundamental, I think - looking after the
> infrastructure is time consuming and for some items like requests and
> security issues, quick responses are also needed. There just isn't
> enough admin time being invested and what work is done lacks
> coordination and long-term focus.
I'll be a bit more specific about the problems I've noticed:
1. It's unclear who own which service.
2. It's unclear when said owner stops taking ownership of
said service.
3. It's unclear who are the key stakeholders in said service.
4. It's unclear which configuration and/or resource files are
key to that service.
5. It's unclear when a service should be retired or has been
superceded.
In other words, the issue is about services: a web service, a
persistent daemon, a file tree, mailing list, or whatever.
Ideally, there would be clear documentation about all those points, a
central repository containing *any* significant configuration files,
and some sort of monitoring so that service owners are notified when
their services are unavailable. (If the documentation approached
superb, any admin would have the knowledge to troubleshoot any
service.)
I've never minded doing basic *system* administration for abbot, but
I've never had the knowledge or documentation to do decent *service*
administration.
--
Paul Heinlein Galois, Inc.
Systems Administrator 421 SW Sixth Avenue, Suite 300
[email protected] Portland, Oregon 97204
+1 503 626-6616 x140 http://corp.galois.com/
_______________________________________________
haskell-infrastructure mailing list
[email protected]
http://community.galois.com/mailman/listinfo/haskell-infrastructure
