>The one thing I can think of, > which would be a possible > problem point for a Dell is – In order to swap > a failed disk without > taking the system down, you need to be able to > identify the failed disk, remove > it and replace it, without shutting down into > BIOS. Traditionally on > Windows or Linux on Dell, I’ve done this via > OpenManage, but I need to > know this exists (or a suitable alternative) for > solaris on Dell.
First, the hardware has to support hot swapping. Second, when you're setting the system up, before anything is using it, and with physical access to both system and console, do some tests, e.g. for each disk, dd if=/dev/rdsk/cXtYdZs2 bs=512 of=/dev/null and see which disk stays lit up steadily. If the lights aren't visible, think up some command that will make just one disk at a time seek a lot, and use a stethoscope. :-) Then label the disks with the names that the OS uses for them. That way, if something dies, you won't need an indicator light or magic software to find it, you'll know which disk physically corresponds to what the OS calls it. Also read the man page on cfgadm; there's a couple of commands to use before removing the disk, and a couple more to use after putting in the replacement. Without having tried I can't be sure, but I think that would do the trick. You get hardware and software from the same place, you have one place to ask such questions, and possibly slicker answers. Not to mention that most Sun servers are built like tanks. Decades ago, I saw a Sun server have a few thousand gallons of leaking cooling water (with various additives) pour on it. After all the wet vacs and mops were gone, we pulled the boards out, borrowed one of the fans the janitors had set up to dry the floors, and set up the fan, boards, and chassis so they'd all get dried out. An hour or so later, we put it all back together, flipped the breakers (with gloves or dry shoes or both), and it worked. One disk in an external JBOD by another vendor had died, that's it. Not guaranteed to work like that I'm sure, and not an experiment I'd care to repeat. But very impressive, IMO. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
