From: rossman@columbia.edu
Date: Tue Nov 02 2004 - 16:30:18 EST
I am up against the clock on the following issue, with a system
very close to going production that has severe performance issues:
I am getting extremely slow disk-to-disk copy performance on
a particular server (a Solaris 8 V440 with an attached D1000,
managed with DiskSuite). There are no disk failure indications
in /var/adm/messages, yet a disk-to-disk copy of a 2GB file takes
about 45 minutes.
I can only assume that the DiskSuite LUNs have been set up in
a very inefficient way, but I need some tools to help me determine
whether this really is the case, and if so, how can I best remedy
the situation.
The most bizarre piece of the puzzle seems to be that I have
three-way stripes, but the middle disk in the three seem not
to be getting any IO (or not nearly as much as the other two).
Anyone care to offer any useful suggestions as to favorite tools,
hints, tricks, tips to tracking down and fixing IO performance
bottlenecks? I'm using things like iostat and sar, but those
are cumbersome, and don't seem to tell me what I really need
to know.
adTHANKSvance,
KR
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:29:39 EDT