Subject: beta software
From: wesley.craig@umich.edu
Date: Wed Oct 02 1996 - 13:07:36 EDT
Just a reminder to other folks of what "beta" means. While we haven't
reproduced this problem here, yet, this is the second report I've had
that netatalk can freeze certain Solaris machines. BTW,
/etc/drivers_{aliases,classes} getting corrupted during driver load is
a well known (to streams developers) Solaris bug.
:wes
------- Forwarded Message
To: wesley.craig@umich.edu
Subject: Re: netatalk 1.4 beta 1
In-reply-to: Your message of "Tue, 01 Oct 1996 18:31:57 EDT."
<199610012231.SAA21871@terminator.rs.itd.umich.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <10101.844248573.1@phaedrus.uchicago.edu>
Date: Wed, 02 Oct 1996 04:29:34 -0500
From: steve farrell <spfarrel@midway.uchicago.edu>
let me start off by saying i think netatalk is really cool, and i've
had great success with 1.3.3 under linux.
now, i have a story for you:
i was really excited to see the new release, and i immediately grabbed
it and prepared to install it on our SS4:
SunOS meno 5.5 Generic sun4m sparc SUNW,SPARCstation-4
i'm using the sun c compiler:
cc: SC4.0 18 Oct 1995 C 4.0
being a little over anxious and not reading the Makefiles fully,
i went to the sys/solaris directory ran make, make install, make
kinstall from there. i got a bunch of errors, and after a while
figured out i was supposed to run make from the top of
the build directory.
the visible damage was that it put netatalk dirs in /atalk, not
/usr/local/atalk, which seemed moderately harmless.
however, when i ran the script (/etc/init.d/atalk start), the system
hard-locked up. frozen. we bought this machine in march and it's been
up ever since i plugged it in, so this came as a shock.
evidentally it did not sync and the files
/etc/drivers_aliases
/etc/drivers_classes
were trashed, as well as the atalk script never seemed to make it to
disk.
i'm not a solaris expert (it's always just *worked* for me), so it
was quite a shock to see it get about 2 seconds into the boot process
before getting a screenful of TRAP ERROR blah blah blah and then
a spontaneous reboot.
anyway, since i located the aforementioned damaged files (i had a ZIP
drive set up as a backup boot in case something like this happens =),
i was able to restore the system... but not after a good scare; this
machine is quite critical to my job and i shouldn't have been putting
beta drivers in the kernel in the first place. i also removed all
traces in /usr/kernel/strmod and /usr/kernel/drv of the ddp, and
the installation in general.
>This message announces netatalk 1.4 beta 1. It's in ~ftp/unix/netatalk
>on terminator.rs.itd.umich.edu. It includes support for SunOS, Ultrix,
>Linux, Solaris and FreeBSD. Enjoy.
well, i can't say i've Enjoy'd so far.
needless to say, i'm writing to you to make sure you're aware of
the current ability of the kernel driver to crash a solaris box,
and issue appropriate warnings, etc, if you deem appropriate.
i DO have another solaris box (an SS1 which is a little more expendable,
and what i should have used in the first place...), so if i can be of help
in making sure this doesn't happen in the future to others, let me know.
best of luck --
steve farrell
------- End of Forwarded Message
This archive was generated by hypermail 2b28 : Sat Dec 18 1999 - 16:24:13 EST