[Beowulf] How do people keep track of computers in your cluster(s)?
Robert G. Brown
rgb at phy.duke.edu
Sun Oct 21 13:24:51 PDT 2007
On Sun, 21 Oct 2007, Carsten Aulbert wrote:
> However, I would like to have something where we have something like a
> large table about the hardware in question. In there information like
> * vendor
> * serial number
> * MAC addresses (eth0, eth1,..., IPMI, RAID,...)
> * maybe even firmware versions and serial numbers of exchangeable
> internal hardware (hard disks)
> * basically all physical information of the box
> another table should hold the current setup, i.e. a mapping between the
> hardware and the "logical" setup, e.g.
> Hardware box number #1234 from above table has in the current setup the
> * hostname
> * IP addresses
> * running services
> * ...
> And finally, another table where special problems, like memory errors
> and the like can be entered.
Wulfware can do some of this, but I never thought about putting in all
of it. And it does have a nice logging interface (wulflogger) that
returns the information as a table, and examples of perl scripts that
can be used to parse and repack the data further if you like.
However, wulfware is open source, and it would be really easy to add
some of this to xmlsysd and then to wulfstat or wulflogger output. The
problem is that some of this isn't available in a standard form in
software. How can you read the serial number of a system in software?
Or firmware version? If you know how to go about this with systems
calls it should be easy enough to add.
Disk information could probably be added. Running services COULD be
added, but boy, there could be a lot of them (I assume you mean the
output of e.g. chkconfig --list or the like).
At the moment, at least, it returns things that are fairly easy (well,
not necessarily "easy" but possible:-) to extract from e.g. /proc tables
or systems calls, but not much else. Things that aren't available in
software (or aren't likely to be useful to global cluster
administration) aren't there.
> I know this sounds just like three medium sized SQL tables, but at least
> I wanted to ask what people are using if more than a single person is
> working on the cluster. One person can probably do this with a simple
> text file and a set of papers in a filing cabinet.
> Thanks a lot for any answers
> Beowulf mailing list, Beowulf at beowulf.org
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
Robert G. Brown
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Lulu Bookstore: http://stores.lulu.com/store.php?fAcctID=877977
More information about the Beowulf