[SAGE] diagnosing bottlenecks on Linux

[John Borwick <borwicjh@wfu.edu>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hey, everybody.  I hope you are doing well.

We have a database machine that's constantly working.  It has an average
load of 2-4, 8 (real) CPUs + hyperthreading, SAN-attached storage, 16 GB
RAM, and a gig-E network connection.

I have attempted to use top, vmstat, iostat, and "netstat -s" to figure
out what's going on with this machine.  I'm sure that historical
monitoring, which we're putting in place, will help.

But--how do you find out a machine's bottleneck?  How do you tell if
it's %utilization of the network, disk, etc.?  Or perhaps--how can you
tell what the average wait time is for various types of I/O?  How can
you tell if it's an application issue (e.g. busywait) or a hardware
limitation?   I'm looking primarily for real-time examination tools.

Thanks for your advice!

Yours,
John
- --
~            John Borwick
~        System Administrator
~       Wake Forest University | web  http://www.wfu.edu/~borwicjh
~       Winston-Salem, NC, USA | GPG key ID             0x797326D5

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (FreeBSD)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBrJq0/81sF3lzJtURAmvwAJ9Q5BI8CSts10CoOzcL3OY/Nm2qQQCgqYDJ
TFEyfVt7y6jtO7HUYd0Zg+0=
=rStg
-----END PGP SIGNATURE-----