From: Kevin Metzger (kevin@pmimail.com)
Date: Fri Oct 11 2002 - 10:08:39 EDT
I'm moving ~25 Gig from a Intel Solaris 8 to my new EMC CLARiiON attached to
an E220R also running Solaris 8. The x86 is a Compaq with SCSI RAID and a
100Fdx network connection, the 220's network is fiber 1G/s. The EMC is
attached via fiber channel. Using ufsdump | ufsrestore, ufsdump logging is
reporting 80-90 hours to complete. The network is Cisco switched and there
are no auto-negotiate on these interfaces. Looking at nnd /dev/tcp and ndd
/dev/udp on the x86:
udp_xmit_hiwat 65536
udp_xmit_lowat 32768
udp_recv_hiwat 65536
and
tcp_xmit_hiwat 65536
tcp_xmit_lowat 32768
tcp_recv_hiwat 65536
On the 220R they are:
udp_xmit_hiwat 8192
udp_xmit_lowat 1024
udp_recv_hiwat 8192
tcp_xmit_hiwat 32768
tcp_xmit_lowat 4096
tcp_recv_hiwat 32768
Should I match the 220's udp and tcp values? Do the 220's need increased even
though the interface should be 10 times as fast?
Any input is welcome.
I'll summarize.
--==========================================================================--
Kevin Metzger direct mobile: 614.378.6396 desk: 800.777.3574x2686
Systems Administrator 614.839.0740 fax
Progressive Medical, Inc. kevin@pmimail.com
--==========================================================================--
_______________________________________________
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 - 23:25:05 EDT