I''ve got a machine(remote, but with an ilo, so I have remote console/serial) that doesn''t boot with xen3. Here''s the boot matrix: +---------+---------------+-------+ | Xen | linux | boots | |---------|---------------|-------| | 2.0.6 | 2.6.11.9 | yes | | | 2.6.12.2 | yes | | | 2.6.12.6 | yes | | 3.0.0 | 2.6.12.6 | no | +---------+---------------+-------+ So, xen3 is causing the system to lockup during boot. Here''s the boot log: Attached are the xen3+2.6.12.6 and plain 2.6.12.6 boot logs and configs. Anyone have any ideas as to what might be wrong? Xen is compiled with the defaults; I''ve tried acpi=ht and acpi=off, with no success. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> Anyone have any ideas as to what might be wrong? Xen is > compiled with the defaults; I''ve tried acpi=ht and acpi=off, > with no success.Are you putting acpi=off on the xen command line? (putting it on the linux command line won''t work). Iam _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 8 Dec 2005, Ian Pratt wrote:> > > > Anyone have any ideas as to what might be wrong? Xen is > > compiled with the defaults; I''ve tried acpi=ht and acpi=off, > > with no success. > > Are you putting acpi=off on the xen command line? (putting it on the > linux command line won''t work).yes. kernel /boot/xen-3.0.0.gz com2=115200,8n1 console=com2,vga noreboot=1 acpi=ht module /boot/xen-linux-2.6.12.6-xen xencons=ttyS31 DOMID=0 panic=1 root=/dev/cciss/c0d0p1 rootflags=data=writeback elevator=deadline _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> Attached are the xen3+2.6.12.6 and plain 2.6.12.6 boot logs > and configs.You might want to try putting maxcpus=1 on the Linux command line. If this works, I suspect it might be the nash module loading bug; try upgrading nash on the machine you run mkinitrd on. Ian _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 8 Dec 2005, Ian Pratt wrote:> > Attached are the xen3+2.6.12.6 and plain 2.6.12.6 boot logs > > and configs. > > You might want to try putting maxcpus=1 on the Linux command line. If > this works, I suspect it might be the nash module loading bug; try > upgrading nash on the machine you run mkinitrd on.I''m not using an initrd. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 8 Dec 2005, Adam Heath wrote:> On Thu, 8 Dec 2005, Ian Pratt wrote: > > > > Attached are the xen3+2.6.12.6 and plain 2.6.12.6 boot logs > > > and configs. > > > > You might want to try putting maxcpus=1 on the Linux command line. If > > this works, I suspect it might be the nash module loading bug; try > > upgrading nash on the machine you run mkinitrd on. > > I''m not using an initrd.I just tried maxcpus=1 on both the xen and linux lines; still no go. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 8 Dec 2005, Adam Heath wrote:> On Thu, 8 Dec 2005, Adam Heath wrote: > > > On Thu, 8 Dec 2005, Ian Pratt wrote: > > > > > > Attached are the xen3+2.6.12.6 and plain 2.6.12.6 boot logs > > > > and configs. > > > > > > You might want to try putting maxcpus=1 on the Linux command line. If > > > this works, I suspect it might be the nash module loading bug; try > > > upgrading nash on the machine you run mkinitrd on. > > > > I''m not using an initrd. > > I just tried maxcpus=1 on both the xen and linux lines; still no go.So, I''m stilling having problems with this hardware; however, now it''s with xen 2.0.6, and 2.6.11.9. The symptoms are like memory corruption; bzip2 reported inconsisent checksums; /var/lib/dpkg/diversions was all nulls. Successive logins returned varying syntax errors in /etc/bash_completion, with the last login having no errors. However, memtest86+.bin reported no problems at all. I''m at a loss as to explain this; the exact same hardware has been running a production server for over a year now, without any problems whatsoever. This issue is solely occuring inside xen. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> So, I''m stilling having problems with this hardware; however, > now it''s with xen 2.0.6, and 2.6.11.9. > > The symptoms are like memory corruption; bzip2 reported > inconsisent checksums; /var/lib/dpkg/diversions was all > nulls. Successive logins returned varying syntax errors in > /etc/bash_completion, with the last login having no errors.Try booting the xen3 demo CD and see if that runs stably. If so, try mounting your disks and see if you can md5sum large files reliably. Ian _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Tue, 13 Dec 2005, Ian Pratt wrote:> > So, I''m stilling having problems with this hardware; however, > > now it''s with xen 2.0.6, and 2.6.11.9. > > > > The symptoms are like memory corruption; bzip2 reported > > inconsisent checksums; /var/lib/dpkg/diversions was all > > nulls. Successive logins returned varying syntax errors in > > /etc/bash_completion, with the last login having no errors. > > Try booting the xen3 demo CD and see if that runs stably. If so, try > mounting your disks and see if you can md5sum large files reliably.Ok, there are some very bad problems going on with smp on 2.0.6. I boot xen 2.0.6, linux 2.6.11.9, with nosmp and noht on the xen line. I can run a gunzip | bzip2 -9, on a 5g file, no problems. I reboot with smp(2 real cpus, 4 virtual). I get an error as *soon* as I log in, like this: =adam@gradall:~/code/bug/apache2-2.0.55$ ssh mirror.kbyou Welcome to Knoppix (Kernel 2.4.20-xfs) -bash: /etc/bash_completion: line 917: syntax error near unexpected token `'' -bash: /etc/bash_completion: line 917: `} '' doogie@beta:~$ logout Connection to mirror.kbyou closed. Killed by signal 1. adam@gradall:~/code/bug/apache2-2.0.55$ ssh mirror.kbyou Welcome to Knoppix (Kernel 2.4.20-xfs) -bash: /etc/bash_completion: line 3443: syntax error near unexpected token `'' -bash: /etc/bash_completion: line 3443: ` fi '' doogie@beta:~$ logout Connection to mirror.kbyou closed. Killed by signal 1. adam@gradall:~/code/bug/apache2-2.0.55$ ssh mirror.kbyou Welcome to Knoppix (Kernel 2.4.20-xfs) -bash: /etc/bash_completion: line 3664: syntax error near unexpected token `&'' -bash: /etc/bash_completion: line 3664: `[ -n "${have:-}" ] && complete -F _urpmi_update urpmi.update'' doogie@beta:~$ logout Connection to mirror.kbyou closed. Killed by signal 1. adam@gradall:~/code/bug/apache2-2.0.55$ ssh mirror.kbyou Welcome to Knoppix (Kernel 2.4.20-xfs) doogie@beta:~$ = This system is a xenU instance. Booting a cd is tricky; the box is in california, while I''m in texas. And the client in california is 40 minutes away from the colo. I can try doing something thru the ilo. I can remote-mount the cd(fetching it with bittorrent now to the machine''s neighbor on the lan). Does the cd do graphical stuff? I''d have to disable that for the remote console to work. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Mon, 12 Dec 2005, Adam Heath wrote:> On Tue, 13 Dec 2005, Ian Pratt wrote: > > > > So, I''m stilling having problems with this hardware; however, > > > now it''s with xen 2.0.6, and 2.6.11.9. > > > > > > The symptoms are like memory corruption; bzip2 reported > > > inconsisent checksums; /var/lib/dpkg/diversions was all > > > nulls. Successive logins returned varying syntax errors in > > > /etc/bash_completion, with the last login having no errors. > > > > Try booting the xen3 demo CD and see if that runs stably. If so, try > > mounting your disks and see if you can md5sum large files reliably. > > Ok, there are some very bad problems going on with smp on 2.0.6. > > I boot xen 2.0.6, linux 2.6.11.9, with nosmp and noht on the xen line. I can > run a gunzip | bzip2 -9, on a 5g file, no problems. > > I reboot with smp(2 real cpus, 4 virtual). I get an error as *soon* as I log > in, like this:Hmm. This now looks like it has nothing to do with xen. I booted normal 2.6.12.6, and started to compile a 2.6.14 kernel, with PAE enabled. I''ve gotten 3 random compiler segfaults. So, it most definately looks like flakey hardware. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Tue, 13 Dec 2005, Adam Heath wrote:> Hmm. This now looks like it has nothing to do with xen. > > I booted normal 2.6.12.6, and started to compile a 2.6.14 kernel, with PAE > enabled. I''ve gotten 3 random compiler segfaults. > > So, it most definately looks like flakey hardware.This turned out to be a mismatched VRM(Voltage Regulator Module) and CPU. Once the VRM was replaced, the machine became stable. The box still doesn''t boot xen 3.0.0 unless I turn off apic, however; normal 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and 2.6.14 all work *with* apic, however. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On 22 Dec 2005, at 05:58, Adam Heath wrote:> This turned out to be a mismatched VRM(Voltage Regulator Module) and > CPU. > Once the VRM was replaced, the machine became stable. > > The box still doesn''t boot xen 3.0.0 unless I turn off apic, however; > normal > 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and 2.6.14 all work *with* apic, > however.You really mean apic, or acpi? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
> The box still doesn''t boot xen 3.0.0 unless I turn off apic, > however; normal 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and > 2.6.14 all work *with* apic, however.Check for a BIOS update. This has solved things for quite a few people that have had this problem. Ian _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 22 Dec 2005, Keir Fraser wrote:> > On 22 Dec 2005, at 05:58, Adam Heath wrote: > > > This turned out to be a mismatched VRM(Voltage Regulator Module) and > > CPU. > > Once the VRM was replaced, the machine became stable. > > > > The box still doesn''t boot xen 3.0.0 unless I turn off apic, however; > > normal > > 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and 2.6.14 all work *with* apic, > > however. > > You really mean apic, or acpi?I really mean apic. xen:/home/shared/doogie# grep apic /boot/grub/menu.lst kernel /boot/xen-3.0.0.gz dom0_mem=256M com2=115200,8n1 console=com2,vga noapic kernel /boot/xen-pae-3.0.0.gz dom0_mem=256M com2=115200,8n1 console=com2,vga noreboot noapic _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 22 Dec 2005, Ian Pratt wrote:> > The box still doesn''t boot xen 3.0.0 unless I turn off apic, > > however; normal 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and > > 2.6.14 all work *with* apic, however. > > Check for a BIOS update. This has solved things for quite a few people > that have had this problem.Nope. Due to the mis-matched CPU/VRM module, the motherboard was swapped. It''s dated at least August; much newer. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
On Thu, 22 Dec 2005, Adam Heath wrote:> On Thu, 22 Dec 2005, Ian Pratt wrote: > > > > The box still doesn''t boot xen 3.0.0 unless I turn off apic, > > > however; normal 2.4.28, 2.4.32, 2.6.12.2, 2.6.12.6, and > > > 2.6.14 all work *with* apic, however. > > > > Check for a BIOS update. This has solved things for quite a few people > > that have had this problem. > > Nope. > > Due to the mis-matched CPU/VRM module, the motherboard was swapped. It''s > dated at least August; much newer.Besides, normal linux, both 2.4 and 2.6 work with apic; I see high-number interrupts. It''s only with xen that it doesn''t work. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel