RE: [Fedora-xen] Is my guest hosed?
by Brian Leyton
Wow, my first question to the list, and I stumped all of the experts ;-)
Let's try a slightly different approach.
Does anyone have, or know where I can get a vmlinuz-2.6.17-1.????.fc5xenU
kernel file? I think if I could just find a kernel of that approximate
vintage, I'll be able to at least boot the guest and get it going again.
Thanks!
Brian
> -----Original Message-----
> From: Brian Leyton [mailto:bleyton@cpe-corp.com]
> Sent: Wednesday, March 07, 2007 9:29 AM
> To: fedora-xen(a)redhat.com
> Subject: [Fedora-xen] Is my guest hosed?
>
> I apologize - this is a long story, and I hope I can explain
> it properly.
>
> I'm somewhat new to Xen, but some time back I managed to get
> a working installation going. I used FC5, and had my guest
> (also FC5) working fine, except for some occasional
> unexplained crashes (but that's a story for another time).
>
> Yesterday, my guest crashed, and I could not get it to
> restart. I rebooted the host, without success. I ran a yum
> update on the host, and then Xen would not even start. I
> tried googling the error messages, and the indication I got
> from that was that there was some incompatibility between the
> Xen tools and Xen itself (I think this diagnosis turned out
> to be wrong
> - more on that in a bit).
>
> So I decided to try upgrading the host to FC6 (using yum),
> the theory being that at least FC6 might have compatible
> versions of everything together.
> When I did this, it complained about an old kernel version
> (2.6.17) being incompatible with a couple of packages, so I
> used RPM and removed this old kernel (bad move).
>
> So FC6 installed properly, and after realizing that it was
> booting with the non-xen-enabled kernel, it now works fine.
> I have a feeling that this was my problem all along. Wish I
> had known that before...
>
> In any case, I was able to get Xen started too. But the
> guest would not start up, because the kernel I had deleted
> was the one that this guest needed. I tried changing the
> .cfg file to point to a newer kernel, but though the guest
> would start up, it will not run properly with the newer kernel.
>
> So I find myself stuck. I searched but could not find a
> 2.6.17 xen enabled kernel anywhere, and I can't figure out
> how to fix this guest. It looks like some automatic yum
> updating has updated the kernel on the guest as well (all of
> the kernel files in /boot are 2.6.19), and I can't find the
> 2.6.17 in there. I'm not sure why the updates that took
> place on the guest would not have updated everything to run
> with the 2.6.19 kernel, but it just does not work.
>
> Anyone have an idea of how I can salvage this guest?
>
> Brian
>
> --
> Fedora-xen mailing list
> Fedora-xen(a)redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-xen
>
17 years, 2 months
Is my guest hosed?
by Brian Leyton
I apologize - this is a long story, and I hope I can explain it properly.
I'm somewhat new to Xen, but some time back I managed to get a working
installation going. I used FC5, and had my guest (also FC5) working fine,
except for some occasional unexplained crashes (but that's a story for
another time).
Yesterday, my guest crashed, and I could not get it to restart. I rebooted
the host, without success. I ran a yum update on the host, and then Xen
would not even start. I tried googling the error messages, and the
indication I got from that was that there was some incompatibility between
the Xen tools and Xen itself (I think this diagnosis turned out to be wrong
- more on that in a bit).
So I decided to try upgrading the host to FC6 (using yum), the theory being
that at least FC6 might have compatible versions of everything together.
When I did this, it complained about an old kernel version (2.6.17) being
incompatible with a couple of packages, so I used RPM and removed this old
kernel (bad move).
So FC6 installed properly, and after realizing that it was booting with the
non-xen-enabled kernel, it now works fine. I have a feeling that this was
my problem all along. Wish I had known that before...
In any case, I was able to get Xen started too. But the guest would not
start up, because the kernel I had deleted was the one that this guest
needed. I tried changing the .cfg file to point to a newer kernel, but
though the guest would start up, it will not run properly with the newer
kernel.
So I find myself stuck. I searched but could not find a 2.6.17 xen enabled
kernel anywhere, and I can't figure out how to fix this guest. It looks
like some automatic yum updating has updated the kernel on the guest as well
(all of the kernel files in /boot are 2.6.19), and I can't find the 2.6.17
in there. I'm not sure why the updates that took place on the guest would
not have updated everything to run with the 2.6.19 kernel, but it just does
not work.
Anyone have an idea of how I can salvage this guest?
Brian
17 years, 2 months
Win Xp 64 bit install
by Abhishek Gaurav
Hi,
I am trying to install windows xp 64 bit (trial version) using
virt-install on a fc6 system. The installation seems to stall when the
message in the vnc console says "Setup is starting windows".
The xm dmesg shows:
(XEN) (GUEST: 13) Booting from CD-Rom...
(XEN) (GUEST: 13) unsupported PCI BIOS function 0x0E
(XEN) (GUEST: 13) int13_harddisk: function 15, unmapped device for ELDL=81
(XEN) (GUEST: 13) *** int 15h function AX=E980, BX=00F2 not yet supported!
(XEN) (GUEST: 13) *** int 15h function AX=EC00, BX=0002 not yet supported!
(XEN) Trying to set reserved bit in EFER: 4d01
The last message "Trying to set ... " actually comes up just before it
stalls.
What might be causing this ?
I have also tried to install an older Win Xp home sp1, which actually
goes past the above messages (without "trying to set ...") but ends up
in the famous black screen upon first reboot.
It'll be great to know what might be causing the above problems.
Thanks,
Abhishek.
17 years, 2 months
physical memory is reduced :-(
by John Sanabria
Hi,
i had installed the last updates for FC6 in a machine with 1 Giga of RAM.
The machine's ip is 136.145.116.88.
------------------------------------------------------
[root@dhcp-crl-116-88 ~]# free
total used free shared buffers cached
Mem: 962032 281512 680520 0 16116 191412
-/+ buffers/cache: 73984 888048
Swap: 2031608 0 2031608
[root@dhcp-crl-116-88 ~]# uname -a
Linux dhcp-crl-116-88.ece.uprm.edu 2.6.19-1.2911.6.4.fc6xen #1 SMP Sat Feb
24 14:57:17 EST 2007 i686 i686 i386 GNU/Linux
------------------------------------------------------
now, i execute a "xm create xentemplate_6 -c", and voila! i lost approx
250 Mbytes of PHYSICAL RAM, :-(
------------------------------------------------------
[root@dhcp-crl-116-88 ~]# free
total used free shared buffers cached
Mem: -> 764928 <- 375796 389132 0 17536 281980
-/+ buffers/cache: 76280 688648
Swap: 2031608 0 2031608
------------------------------------------------------
then, shutdown the virtual machine, and i did not get back the lost 250
Mbytes of RAM.
anybody knows how to get back my physical memory after to shutdown the
virtual machine? is that behavior a bug?
if i run two virtual machines, my physical memory is reduced until 512 MB
of Physical RAM. At this time, the only way to get 1Gbyte of RAM is
rebooting the machine :-(.
Thanks a lot for your suggestions.
http://ece.uprm.edu/~s047267
http://del.icio.us/josanabr
http://blog-grid.blogspot.com
17 years, 2 months
RE: [Fedora-xen] Kickstart slowness with virt-install
by Hanks, Dan
> -----Original Message-----
> From: Daniel P. Berrange [mailto:berrange@redhat.com]
> Sent: Wednesday, March 07, 2007 2:54 PM
> To: Hanks, Dan
> Cc: fedora-xen(a)redhat.com
> Subject: Re: [Fedora-xen] Kickstart slowness with virt-install
>
> On Wed, Mar 07, 2007 at 01:50:43PM -0800, Hanks, Dan wrote:
> > Hi folks,
> >
> >
> >
> > I've been able to kickstart a number of VMs using the virt-install
tool,
> > and so far have had pretty good success. One aspect of these
installs
> > has me a bit concerned, though. In most cases, anytime the kickstart
> > needs to do intensive disk activity (such as formatting partitions,
or
> > installing all the rpms) there are noticible hangs, which I'm
guessing
> > come from some kind of IO wait. The result is that a kickstart which
> > should take < 10 minutes ends up taking a half-hour or so.
>
> What kind of virtual disk image are you using for the guest ? A
partition
> or a file - if the latter is it sparse, or non-sparse. Basically
sparse
> files will be horribly slow because every time the host OS has to
extend
> the sparse file to allocate real blocks it needs to do a journal sync
on
> the host FS. This destroys performance of I/O from the guest until the
> sparse file is fully-allocated.
I'm using files. I've been using a command-line such as this for the
install:
virt-install -m "00:16:3e:00:00:01" -n hostname -r 500 --vcpus=2 -f
/var/lib/xen/images/myhost.xen.img -s 4 --nographics -p -l
ftp://host/fc6/distro/i386/os -x "ks=http://kshost/ks.cfg"
Adding --nonsparse into the args looks like it fixes the speed issues
(albeit adds a bit of extra time right up front to allocate the entire
disk image). I imagine using partitions instead of files would be faster
altogether. I'll have to explore that option (files are just so nice to
ship around when needed).
The verbiage for the virt-install command-line args is (to me) a bit
confusing in this matter:
"-- nonsparse Don't use sparse files for disks. Note that this
will be significantly slower for guest creation"
What does the "slower" here refer to? Slower up-front time to create the
image file? Yes, but the actual install will be faster since it wont
have to keep allocating more space for the disk image. Thoughts?
Thanks for your help--this was exactly the piece of information I
needed,
-- Dan
17 years, 2 months
Kickstart slowness with virt-install
by Hanks, Dan
Hi folks,
I've been able to kickstart a number of VMs using the virt-install tool,
and so far have had pretty good success. One aspect of these installs
has me a bit concerned, though. In most cases, anytime the kickstart
needs to do intensive disk activity (such as formatting partitions, or
installing all the rpms) there are noticible hangs, which I'm guessing
come from some kind of IO wait. The result is that a kickstart which
should take < 10 minutes ends up taking a half-hour or so.
I'm using pretty standard hardware, Dell 1850s with 2 internal drives.
Xvda setup to be 4G, using 500M of memory. Pretty vanilla installs
really.
Any insight into what might be causing these IO hangs?
Thanks,
-- Dan
17 years, 2 months
Time in an FC7 QEMU quest
by William Murray
I assume this is the correct list for QEMU?
I made a Scientific Linux (~RH-EL 4) QEMU guest, which works well, but
the time runs about a factor of two fast. Connecting to timeservers
somehow doesn't help.
Does anyone know how I can address this?
Thank you,
Bill
--
__________________________________________________________
Bill Murray w.j.murray(a)rl.ac.uk (41)-1235-446256
17 years, 2 months
ssh to QEMU guest
by William Murray
I assume this is the correct list for QEMU?
I would like to SSH to my QEMU quest (FC7 T2 host) and I cannot work
out how to do it. I found some suggestions about TUN/TAP on WWW. Is that
the correct lines?
Any pointers?
Thank you,
Bill
--
__________________________________________________________
Bill Murray w.j.murray(a)rl.ac.uk (41)-1235-446256
17 years, 2 months
para or fully virtualization
by Henk Altena
Hi everybody
Following the quickstart guide of xen I can only use para-virtualized
guests because my Asus A8V Deluxe mainboard with AMD 64 bit processor
does not have the svm flag.
Can anyone tell me the difference between Para- and Fully-virualized
guests ?
What are the limitations of a system that only supports para-virtulazed
guests ?
Henk
17 years, 2 months