Xen users - Aug 2008 - Xen Fedora - http runs away and takes VM with it?

I''m a xen newbie and not very deeply experienced with linux (years of
living in a Solaris bubble) and I have a situation that''s just
puzzling me.
The Xen Host is a Sun v20z running RHEL5. There are two guests running
paravirtualized Fedora 9 (the v20x being too old to support HVM).
Everything''s pretty much updated.

The first guest works fine. In particular, I have a LAMP server *running*.

On the second guest, which I''ve both built standalone and built as a
clone of the first guest, when I start http it hangs with the CPU 49%
busy - hangs to the point where I can''t pause or shutdown the guest,
the xm commands don''t work, and I have to reboot the host (dom0).
It doesn''t matter if the first guest is running or shutdown, and more
memory didn''t change it. But the second guest is a clone of the first!
I even built a new clone with a new name and new IP.  It''s a clone,
dangit, why doesn''t it work the same?

(in cloning, I edited /etc/hosts, and changed the IP and mac address
in /etc/sysconfig/network-scripts/ifcfg-eth0  and
/etc/sysconfig/network, and edited httpd.conf, and that''s it, maybe I
missed something?)

Running strace on "service httpd start" is very unhelpful. I see it
call /etc/init.d.httpd , do some stating of /sbin/env and /bin/env,
starts a child, then :
rt_sigactionm(SIGINT,{0x807c580, [], 0}, {SIG_DFL}, 8) = 0
waitpid(-1, Starting httpd:

and that''s all, it freezes mid-line, CPU pegging at roughly 49%, until
I reboot *dom0* to regain control.

Any thoughts on what to try next very welcome!


thanks Betsy
(who should have had this ready to hand over a long time ago!)

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

I''m a xen newbie and not very deeply experienced with linux (years of
living in a Solaris bubble) and I have a situation that''s just
puzzling me.
The Xen Host is a Sun v20z running RHEL5. There are two guests running
paravirtualized Fedora 9 (the v20x being too old to support HVM).
Everything''s pretty much updated.

The first guest works fine. In particular, I have a LAMP server *running*.

On the second guest, which I''ve both built standalone and built as a
clone of the first guest, when I start http it hangs with the CPU 49%
busy - hangs to the point where I can''t pause or shutdown the guest,
the xm commands don''t work, and I have to reboot the host (dom0).
It doesn''t matter if the first guest is running or shutdown, and more
memory didn''t change it. But the second guest is a clone of the first!
I even built a new clone with a new name and new IP.  It''s a clone,
dangit, why doesn''t it work the same?

(in cloning, I edited /etc/hosts, and changed the IP and mac address
in /etc/sysconfig/network-scripts/ifcfg-eth0  and
/etc/sysconfig/network, and edited httpd.conf, and that''s it, maybe I
missed something?)

Running strace on "service httpd start" is very unhelpful. I see it
call /etc/init.d.httpd , do some stating of /sbin/env and /bin/env,
starts a child, then :
rt_sigactionm(SIGINT,{0x807c580, [], 0}, {SIG_DFL}, 8) = 0
waitpid(-1, Starting httpd:

and that''s all, it freezes mid-line, CPU pegging at roughly 49%, until
I reboot *dom0* to regain control.

Any thoughts on what to try next very welcome!


thanks Betsy
(who should have had this ready to hand over a long time ago!)

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users