Return to BSD News archive
Newsgroups: comp.unix.bsd.netbsd.misc,comp.unix.bsd.bsdi.misc,comp.unix.solaris,comp.unix.aix
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!paladin.american.edu!gatech!newsfeed.internetmci.com!in1.uu.net!deshaw.com!christos
From: christos@deshaw.com (Christos Zoulas)
Subject: Re: ISP hardware/software choices (performance comparison)
Message-ID: <DL86Iy.DKD@deshaw.com>
Sender: usenet@deshaw.com
Nntp-Posting-Host: es2
Organization: D. E. Shaw & Co.
References: <4d4153$437@durban.vector.co.za> <4d8m52$sv5@sundog.tiac.net> <4d9cu6$cnk@mail.fwi.uva.nl>
Distribution: inet
Date: Mon, 15 Jan 1996 13:52:10 GMT
Lines: 53
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.netbsd.misc:1969 comp.unix.bsd.bsdi.misc:2115 comp.unix.solaris:57270 comp.unix.aix:68741
In article <4d9cu6$cnk@mail.fwi.uva.nl> casper@fwi.uva.nl (Casper H.S. Dik) writes:
>rmk@rmkhome.com (Rick Kelly) writes:
>
>>Solaris 2.x: plug it into an SBUS slot, then "boot -r", then watch as
>>all your devices get reassigned and nothing mounts anymore.
>
>Nope, only devices you move get reassigned.
>
>The newly detected card will get the next free number for the specific
>controller type.
>
>>It is only easy if you always put whatever new card your installing as
>>the last card on the bus. This can be especially confusing with the
>>sparc 1000 and sparc 2000.
>
>That's not true, unless you move othe cards around.
>
>The /etc/path_to_inst file remembers which controllers you had before and
>they will not be renumbered.
>
>If you reinstall the machine, then your controllers may get different
>numbers.
>
Oh please, don't talk to me about the gazillion files Solaris needs to
boot cleanly. One of the people working for me moved
/etc/path_to_inst{,.old}. After that when we rebooted the machine, we
were asked if we wanted that file to be recreated: Answering yes, just
continued booting without fixing the file! Then the machine would come
up, but we had no access to the devices to fix the problem. Answering
no just rebooted (now here is for a useful option?!?).
The funniest part was that this was the spare root of the machine which
we had to use after the main raid root became unbootable when we
de-installed and re-installed a device driver. This happened because of
another bug in Solaris-2.5 corrupting /etc/driver_aliases when a device
is removed and re-installed.
And finally, the thing was the most hillarious was that we tried to boot
diskless Solaris from SunOS to fix it (since we don't have any other
Solaris machine at that site). This for some reason did not work. Finally,
we booted the machine diskless using SunOS, mounted the Solaris drive
under SunOS and did 'mv /etc/path_to_inst{.old,}'...
I am not blaiming Solaris for any of this; we are novices in Solaris and
it could be that an experienced Solaris user can fix all of this. I just
miss the old days when all you needed to boot was bootblocks, a good
kernel and init.
christos
----
All of these are my own opinions and not the ones of my employer...