Using kstat for ce interfaces
Since I don’t seem to remember the kstat options for checking the link status of a ce interface, I’m putting it here.
# kstat -p ce:::link_up
will produce
ce:0:ce0:link_up 0 ce:1:ce1:link_up 1 ce:2:ce2:link_up 1 ce:3:ce3:link_up 0 ce:4:ce4:link_up 0 ce:5:ce5:link_up 0
It can even do regular expressions, so you can do something like this to see your speed, duplex and link status:
# kstat -p 'ce:0::/link_(speed|up|duplex)/' ce:0:ce0:link_duplex 2 ce:0:ce0:link_speed 100 ce:0:ce0:link_up 0
Tuesday January 4, 2005 · Permalink
Setting interface speeds on Gigaswift ce interfaces with ce.conf
There is (unfortunately) not much information about correctly setting interface speeds on Solaris systems with their Gigaswift (ce) cards using the ce.conf file instead of using ndd commands after the interface has been brought up. Even some of the documentation on the Gigaswift device driver was incomplete/incorrect (don’t ask me how I know this).
The one page that seems to be complete and correct is their free infodoc 72033: Correct Use of the Driver Configuration File to Force the Sun™ GigaSwift Ethernet Adapter’s Operational Parameters. Hopefully a few more links to the page will help make it just a little easier to find.
Friday August 13, 2004 · Permalink
Bad day for a SunFire 12K
Woah, this SunFire 12K certainly had a bad day due to the courier company’s mistake.
Thanks Azeem Jiva, via PlanetSun for the link.
Thursday July 22, 2004 · Permalink
Correctable Memory Errors in Sun Hardware
Notes from the PDF document Soft Memory Errors and Their Effect On SunFire Systems published April 2002.
There are three categories of memory errors, intermittent, persistent and sticky. Intermittent means that the second read of the data produced the right result. Persistent means the second read produced the same incorrect, but correctable result. Sticky means that the bit could not be changed to the correct value after the testing of the persistent error.
Sun recommends that memory experiencing any sticky errors or three or more persistent errors on the same DIMM be replaced.
Wednesday May 21, 2003 · Permalink