Hi, I''m tryng to set up Xen, I have errors and google does not help me. The situation is simple: I installed Xen-3.0.2 on gentoo. The Dom0 kernel is up and running, all works perfectly. But when I try to start a DomU, xm raises this error: # xm create -c <conf_file> Using config file <conf_file> Error: <ProtocolError: RPC2/: -1> It''s all. After, `xm list` shows that the DomU is in "pause status". If I try to unpause It (`xm unpause`) the DomU''s status becomes "-----". But DomU does not start. Log shows an error while loading the DomU: ... [2006-07-06 18:50:10 xend] CRITICAL (SrvDaemon:190) Xend died due to signal 6! Restarting it. [2006-07-06 18:50:10 xend] INFO (SrvDaemon:278) Xend Daemon started [2006-07-06 18:50:10 xend] INFO (SrvDaemon:282) Xend changeset: Sat Apr 8 12:14:27 2006 +0100 9598:1bce05ff1e52. ... The complete Log-Session is attached to this mail. Best Regards -- efphe _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
efphe wrote:> I''m tryng to set up Xen, I have errors and google does not help me. > > The situation is simple: I installed Xen-3.0.2 on gentoo.By following the instructions at http://www.gentoo-wiki.com/Xen ? Aside from the xen-related ebuilds and their dependencies, you''re running a stable up-to-date Gentoo system?> The Dom0 kernel is up and running, all works perfectly. > But when I try to start a DomU, xm raises this error: > > # xm create -c <conf_file> > Using config file <conf_file> > Error: <ProtocolError: RPC2/: -1>> Log shows an error while loading the DomU: > > ... > [2006-07-06 18:50:10 xend] CRITICAL (SrvDaemon:190) Xend died due to signal 6! Restarting it.You need to find out why Xend is being terminated by the kernel (ie. signal 6). Make sure you have "(loglevel DEBUG)" in /etc/xen/xend-config.sxp and check out /var/log/{xend.log,xend-debug.log} You may need to remerge xen with the debug USE-flag set, so debugging symbols will be available in any core dumps produced by xend. Don''t forget to enable core dumps by running "ulimit -c unlimted" prior to starting Xen. If necessary, stop xend from autostarting at boot ("rc-update del xend default"). Once xend has been terminated with signal 6 it should produce a core dump, which you may have to search for using something like "find / -name core -type f" - you might like to read http://www.linuxhowtos.org/Tips%20and%20Tricks/coredump.htm to see how the name of core dumps can be changed. Once you''ve located the core dump run "file path-to-core-dump" to see what program caused it, then load the program and core dump into gdb (which you''ll need to emerge, if you don''t already have it) with "gdb path-to-program path-to-core". At the prompt type "backtrace", which should help identify exactly where in xend the crash occurred. These instructions are C-specific, so may not be as useful for a python-based program, such as xend. Hopefully, a xen hacker can weigh in on this. Cheers Andrew _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> Another difference is the USE flag "hardened", that I used to > build the entire system. Now, I''m going to rebuild world without that > flag. Tomorrow I''ll see if that is the question.I came across the same problem when running a hardened Gentoo system. I was able to work around the problem by recompiling both python and xen-tools with the hardened-nossp use doing the following: gcc-config x86_64-pc-linux-gnu-3.4.4-hardenednossp emerge -va --oneshot python xen-tools gcc-config x86_64-pc-linux-gnu-3.4.4 HTH, Brad _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Fri, 07 Jul 2006 10:15:50 +1000 Andrew Ross wrote: |By following the instructions at http://www.gentoo-wiki.com/Xen ? Exactly. But in this HowTo the version is 3.0.0. The only ebuild present in portage, today, is the 3.0.2. Another difference is the USE flag "hardened", that I used to build the entire system. Now, I''m going to rebuild world without that flag. Tomorrow I''ll see if that is the question. I''ll try another solution too: I''ll install xen without "hardened" and "ntplonly" flags on another host. |Aside from the xen-related ebuilds and their dependencies, you''re |running a stable up-to-date Gentoo system? Yes. | |> The Dom0 kernel is up and running, all works perfectly. |> But when I try to start a DomU, xm raises this error: |> |> # xm create -c <conf_file> |> Using config file <conf_file> |> Error: <ProtocolError: RPC2/: -1> | |> Log shows an error while loading the DomU: |> |> ... |> [2006-07-06 18:50:10 xend] CRITICAL (SrvDaemon:190) Xend died due to signal 6! Restarting it. | |You need to find out why Xend is being terminated by the kernel (ie. |signal 6). ..... |Once you''ve located the core dump run "file path-to-core-dump" to see |what program caused it, then load the program and core dump into gdb |(which you''ll need to emerge, if you don''t already have it) with "gdb |path-to-program path-to-core". At the prompt type "backtrace", which |should help identify exactly where in xend the crash occurred. Yes, me too I''m a developer, but now ... no time to debug :/ Do you have direct experience with gentoo and xen? br -- efphe _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
efphe wrote:> On Fri, 07 Jul 2006 10:15:50 +1000 > Andrew Ross wrote: > > |By following the instructions at http://www.gentoo-wiki.com/Xen ? > > Exactly. But in this HowTo the version is 3.0.0. > The only ebuild present in portage, today, is the 3.0.2.The HOWTO was updated for 3.0.2 a while ago, but one of the grub entries still says 3.0.0 - that is merely a cosmetic difference, and in no way affects the HOWTO. I''ve fixed it now.> Another difference is the USE flag "hardened", that I used to > build the entire system. Now, I''m going to rebuild world without that > flag. Tomorrow I''ll see if that is the question.Hopefully Brad''s answer will be of assistance in that regard. My dom0 isn''t hardened.> Do you have direct experience with gentoo and xen?Yes, but not in debugging - only as a user. I''ve attached my "emerge --info" from a functioning Xen box, but I suspect you''ll be fine with Brad''s hardened hint. aross@oak ~ $ cat /etc/portage/package.keywords app-emulation/xen ~x86 app-emulation/xen-tools ~x86 sys-devel/dev86 ~x86 sys-kernel/xen-sources ~x86 aross@oak ~ $ cat /etc/portage/package.use sys-libs/glibc nptlonly app-emulation/xen-tools screen Cheers Andrew _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Fri, 07 Jul 2006 11:27:04 +1000 Brad Plant wrote:> > Another difference is the USE flag "hardened", that I used to > > build the entire system. Now, I''m going to rebuild world without > > that flag. Tomorrow I''ll see if that is the question. > > I came across the same problem when running a hardened Gentoo system. > I was able to work around the problem by recompiling both python and > xen-tools with the hardened-nossp use doing the following: > > gcc-config x86_64-pc-linux-gnu-3.4.4-hardenednossp > emerge -va --oneshot python xen-tools > gcc-config x86_64-pc-linux-gnu-3.4.4Resolved! I removed the "hardened" USE flag, make.profile now points to default/2006. I Emerged system and world. Now Xen is fine. BTW, this is a dirty problem: maybe we have to signl this behavior. Thanks -- efphe _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users