Dave,
Jim Dunham wrote:
Jake,
I have a pretty fast PC filer (OpenSolaris 11/08) that's
hosting around fifteen iscsi targets for several of our production
services. Some of these targets occasionally begin to read and
write so much that they're hogging the network and/or disk
subsystem, resulting in unacceptable performance for the other
services. Is there a clever way to determine how much of these
resources the targets are consuming and also perform some rate
limiting? Thus far I've been using iostat and manually tracking
down and killing the offenders, but I was hoping there'd be a
proper cli program, most preferably one that could also be
interrogated and frobbed using snmp.
For OpenSolaris 11/08, you are using the iSCSI Target Daemon, and
thus a means to look inside the daemon as to what is doing can be
accomplished by DTrace: http://www.solarisinternals.com/wiki/index.php/DTrace_Topics_iSCSI
Once you have observe-ability into the iSCSI Target, one can then
use CrossBow, another part of OpenSolaris to manage TCP/IP QoS
(Quality of Service): http://opensolaris.org/os/project/crossbow
Is there a COMSTAR/itadm equivalent of 'iscsitadm show stats'
equivalent for snv_112+? That was very handy.
No, there is no 'stat' command available for the iSCSI Target in
COMSTAR.
Development work is underway across all of COMSTAR to provide DTrace
observation across the COMSTAR LUN provider (sbdadm) and each of the
different port providers (FC, iSCSI, iSER, and soon SRP).
http://wikis.sun.com/display/DTrace/Providers
- Jim
--
Dave
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss
