Date: Tue, 13 Aug 2002 23:05:23 +0200 (CEST) From: "Hartmann, O." <ohartman@klima.physik.uni-mainz.de> To: freebsd-stable@freebsd.org Subject: Still fxp/nfsd problems after today's recent cvsupdate Message-ID: <20020813221318.U384-100000@klima.physik.uni-mainz.de>
next in thread | raw e-mail | index | archive | help
Dear Sirs. Within tha last few days I ran several times into problems due to a bug in FreeBSD 4.6-STABLE. The last full cvsupdate has been done yesterday evening, but after a make world, mergemaster and installation of the appropriate kernels, it took several hours when the first problems occured again. Our servers and diskless terminals use all Intel EtherExpress Pro NICs and so they suffered from the bug as mentioned here in the mailing list. Yesterday I could fix the problem by cvsupdating the sources 'back' to 2002.08.07.12.00.00 and recompiling a kernel. Today I'm not able to solve the problem by this 'nasty' trick. I did the last cvsupdate today at 1300 CEST. A make world, mergemaster, make buildkernel worked well and after rebooting all of our servers again they did not show these strange problems of the last days. But now none of our diskless terminals is willing to boot. After loading, booting teh kernel each terminal tries to work its scripst but when mounting the file systems from /etc/fstab they abort. The last message I see is the following: +++ create lastlog . . . Warning: Block size restricts cylinders per group to 25 /tmp/dev.cpio.gz: No such file or directory cpio: Memory exhausted Mounting /etc/fstab filesystems failed, startup aborted We use the latest rc.diskless scripts, so I did not realize a change since the last cvsupdate yesterday and the most recent today. It seems that the kernel on each terminal run out of memory (most terminals have only 64 MB, but others have 128MB) and therefore it can not mount the NFS file systems. I have no glue whats going on nor an idea how to track down the problem. It is courious that it worked yesterday and today not. As mentioned, yesterday it was possible to fix these problems by bulding a kernel with older bug-free sources, but today it seems to be deeper inside of a module? Well, could this have to do something with the bug in the fxp driver mentioned here? Well, if this is a stupid config mistake (but when did essential things change?), please send me an email or a tip, how to track down this problem. Thanks a lot. As attachment, there is the dmesg output of the boot server (NFS boot process). But it is truncated (system is now up for 1 hour!). Strange! Boot server and terminals use polling (but polling disabling does not fix the problem). ---- dmesg 180.54:25 in via fxp0 -- MfG O. Hartmann ohartman@klima.physik.uni-mainz.de ------------------------------------------------------------------ IT-Administration des Institutes fuer Physik der Atmosphaere (IPA) ------------------------------------------------------------------ Johannes Gutenberg Universitaet Mainz Becherweg 21 55099 Mainz Tel: +496131/3924662 (Maschinenraum) Tel: +496131/3924144 (Buero) FAX: +496131/3923532 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020813221318.U384-100000>