[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] xen-2.0 20040910 problems
Ian Pratt wrote: The previous snapshot - 20040906 - worked ok but encountered "Kernel panic: Unable to reduce memory reservation ..." as discussed elsewhere in this list.I have been experimenting with xen 2.0 and before that with xeno-unstable. It has proved interesting and promising However, the last few updates have been problematic. In particular it looks as if the ext3 filesystem is not correctly handled.Can you give us an idea when it last worked reliably for you? At present the snapshot for 20040911 seems to show no signs of the filesystem problem, so I probably was seeing a side effect of some of the changes you were making in the pagetable system.Recent history: 20040907: occasional kernel BUGs 20040908: instant reboot loop when booting xen 20040909: mysterious filesystem problems 20040910: mysterious filesystem problemsI suspect that these are all a result of experiencing some virtual memory bug. However I have seen at least one instance of the "Kernel panic: Unable to reduce memory reservation ..." problem. CPU: AMD Athlon(tm) XP 2400+ stepping 01 ('uname -a' shows architecture as i686). 512Mbytes RAM. El-cheapo-self-build(tm). The Xen kernels are built straight out of the snapshot with no local configuration.It's pretty odd, as we haven't seen any problems in our testing, but its possible that our automated testing doesn't run enough concurrent domains to tweak it. Can you tell us a bit more about your setup, e.g. SMP? Xeon orOpteron? Nothing much (no exerciser running). But running sshd, postfix, apache, proftpd, webmin, xinetd. Each xenU domain has only 32Mbytes, so the virtual memory would have been fairly active. Each domain would like to have a swap partition but doesn't get one as yet largely because I haven't got round to changing the configuration from the minimum required to make untrusted domains boot and run. But I'm not clear how swap and xen are intended to interact, and whether I should configure swap devices for untrusted domains.What are the other domains doing when 'rpm -qa' explodes? What version of libc are you using i.e. 'ldd /bin/rpm' ? (We've never used mandrake, but I doubt this is significant) ....$ ldd `which rpm` linux-gate.so.1 => (0xfbffd000) librpm-4.2.so => /usr/lib/librpm-4.2.so (0x40029000) librpmdb-4.2.so => /usr/lib/librpmdb-4.2.so (0x40082000) librpmio-4.2.so => /usr/lib/librpmio-4.2.so (0x4016d000) libpopt.so.0 => /lib/libpopt.so.0 (0x401af000) libelf.so.1 => /usr/lib/libelf.so.1 (0x401b7000) libbeecrypt.so.6 => /usr/lib/libbeecrypt.so.6 (0x401c9000) librt.so.1 => /lib/i686/librt.so.1 (0x401e7000) libpthread.so.0 => /lib/i686/libpthread.so.0 (0x401fa000) libbz2.so.1 => /usr/lib/libbz2.so.1 (0x4024c000) libc.so.6 => /lib/i686/libc.so.6 (0x4025b000) /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)I very much doubt if use of Mandrake has much to do with it. My main issue with Mandrake is that the standard boot files need quite a lot of tweaking if you want to have an nfs root. But this also applies to some other distributions I've tried. If you can narrow down to a simple to reproduce test case I'm surewe can find and fix this. It currently looks to have been a passing glitch. I realised that, but hadn't yet made very serious attempts to find my way into the guts of the python.On a different topic, I have never yet seen the web interface work for long enough to be useful - it always crashes with the int-text type clash that was mentioned somewhere in the changelogs as having been fixed.The web interface is very much a work-in-progress that we're hoping other people will contribute on. Thanks Peri ------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement onwho ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/xen-devel
|
![]() |
Lists.xenproject.org is hosted with RackSpace, monitoring our |