Hi, I have already set up a VM that can access the network using the NAT mode. The problem I have is that I''d like to create another VM that also has access to the network. The problem I get is that when a VM is started, the other one will refuse to start. Actually it starts, but when I want to "xm console" into it I get the following error message: "xenconsole: Could not read tty from store: No such file or directory". Where could this problem come from? I don''t think it is a problem with the network, because I deactivated network support in one of the VM and it still refuses to start. Both VM work when there are launched alone, the issue is that when a VM is already running, the other can''t start, this issue happens for both VMs. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi, I have already set up a VM that can access the network using the NAT mode. The problem I have is that I''d like to create another VM that also has access to the network. The problem I get is that when a VM is started, the other one will refuse to start. Actually it starts, but when I want to "xm console" into it I get the following error message: "xenconsole: Could not read tty from store: No such file or directory". Where could this problem come from? I don''t think it is a problem with the network, because I deactivated network support in one of the VM and it still refuses to start. Both VM work when there are launched alone, the issue is that when a VM is already running, the other can''t start, this issue happens for both VMs. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
>From a buggy version of Xen perhaps. I''ve seen this on numerousCentOS 5.1 boxes (not 5.0), the only workaround I found was to restart xend. FYI, it was also characterized by a loss of routed network connectivity on several VMs, which is how I''d find out the problem existed to begin with. -Ray On Sat, Jun 7, 2008 at 2:04 PM, Thomas <iamkenzo@gmail.com> wrote:> Hi, > > I have already set up a VM that can access the network using the NAT mode. > > The problem I have is that I''d like to create another VM that also has > access to the network. The problem I get is that when a VM is started, > the other one will refuse to start. Actually it starts, but when I > want to "xm console" into it I get the following error message: > "xenconsole: Could not read tty from store: No such file or > directory". > > Where could this problem come from? > > I don''t think it is a problem with the network, because I deactivated > network support in one of the VM and it still refuses to start. Both > VM work when there are launched alone, the issue is that when a VM is > already running, the other can''t start, this issue happens for both > VMs. > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Now that I think about it, that error can also come about if you try to console to an HVM (hardware virtual machine) domain. I''m going to assume for the moment that''s not your problem because you made it sound like the second domU is close in configuration to the first one - if true, surely the first one is paravirtualized. I''m CCing the list so they''ll see your direct reply with pertinent information below. -Ray On Sat, Jun 7, 2008 at 2:56 PM, Thomas <iamkenzo@gmail.com> wrote:> Thanks Ray, > > Using "xm info" I got the following info: > > release : 2.6.21.7 > xen_caps : xen-3.0-x86_64 xen-3.0-x86_32p > > I am using a Debian Etch with Xen preinstalled distribution that is > provided by my hosting company. I''ll give them some feedback. >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Thank you Ray, Yeah I basically duplicated my first VM, I simply edited the config file to change the name of the domain, and changed the IP addresses too. Running xentop command gives me this: (I hope the formatting is readable on your side of the email) --------------------------- xentop - 21:22:32 Xen 3.2.1 3 domains: 2 running, 1 blocked, 0 paused, 0 crashed, 0 dying, 0 shutdown Mem: 2088188k total, 2076324k used, 11864k free CPUs: 2 @ 3000MHz NAME STATE CPU(sec) CPU(%) MEM(k) Domain-0 -----r 4041 0.4 1508352 node01 --b--- 6 0.0 262144 node02 -----r 534 100.0 262144 --------------------------- We clearly see that node02 domain is having some trouble to startup. Some other folks proposed that it might be something to add in /etc/inittab to have access to the console, he thinks it has to do with xvc0. PS: sorry Ray, I didn''t notice that the "reply-to" of the emails default to the user that answers the email. It should default to xen-users@lists.xensource.com instead. Can someone change that behavior in the mailing-list? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-07 21:39]:> Yeah I basically duplicated my first VM, I simply edited the config > file to change the name of the domain, and changed the IP addresses > too.Can you provide me one or preferably both configuration files? I might then be able to run some tests on monday. I guess the server is a "Kimsufi XL"? Best regards, Felix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Saturday June 07 2008 03:36:58 pm Thomas wrote:> Yeah I basically duplicated my first VM, I simply edited the config > file to change the name of the domain, and changed the IP addresses > too.I hope you changed the MAC address in your vif= line as well. Don''t have a MAC parm? It''s a good idea - some distros keep renaming eth0 in the domu to eth1, eth2, etc. with each reboot. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Just to be sure, You''re using different disk images / lvm partitions / ... for the different domains ? Thomas schrieb:> Thank you Ray, > > Yeah I basically duplicated my first VM, I simply edited the config > file to change the name of the domain, and changed the IP addresses > too. > > Running xentop command gives me this: (I hope the formatting is > readable on your side of the email) > --------------------------- > xentop - 21:22:32 Xen 3.2.1 > 3 domains: 2 running, 1 blocked, 0 paused, 0 crashed, 0 dying, 0 shutdown > Mem: 2088188k total, 2076324k used, 11864k free CPUs: 2 @ 3000MHz > NAME STATE CPU(sec) CPU(%) MEM(k) > Domain-0 -----r 4041 0.4 1508352 > node01 --b--- 6 0.0 262144 > node02 -----r 534 100.0 262144 > --------------------------- > > We clearly see that node02 domain is having some trouble to startup. > > Some other folks proposed that it might be something to add in > /etc/inittab to have access to the console, he thinks it has to do > with xvc0. > > > PS: sorry Ray, I didn''t notice that the "reply-to" of the emails > default to the user that answers the email. It should default to > xen-users@lists.xensource.com instead. Can someone change that > behavior in the mailing-list? > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users-- Stephan Seitz Senior System Administrator *netz-haut* e.K. multimediale kommunikation zweierweg 22 97074 würzburg fon: +49 931 2876247 fax: +49 931 2876248 web: www.netz-haut.de <http://www.netz-haut.de/> registriergericht: amtsgericht würzburg, hra 5054 _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi, Thank you all for your help. Here is a very long email, with all
the info I thought might be interesting.
@Stephan: I use different lvms for each VM.
@Felix, yes I am using a KimsufiXL server.
@Jim: mac address in vif? I thought it was an IPv4 address I had to
put. What should I put then? Any figures?
Here are my config files:
node01:
--------------------------------------------------------------
# -*- mode: python; -*-
kernel="/boot/domU_32/vmlinuz-2.6.21-xxxx-xenU-ipv4-32"
memory = "256"
maxmem = "512"
name = "node01"
vcpus=1
interface=eth0
dhcp="off"
vif = [ ''ip=10.0.0.1'' ] # replace with DomU''s
IP
ip = "10.0.0.1"
gateway= "10.0.0.128" # Dom0
netmask = "255.255.255.0"
# example for LVM or partition
disk = [ ''phy:/dev/mapper/xenvg-node01,/dev/hda1,w'',
''phy:/dev/mapper/xenvg-node01--swap,hda2,w'' ]
# example for image file
#disk = [ ''file:xen/xenvg-node0001.img,hda1,w'',
''file:/xen/xenvg-node0001-swap.img,hda2,w'' ]
# This uses environment variables XAUTHORITY and DISPLAY. You
# can override that:
# vfb = [
''type=sdl,xauthority=/home/bozo/.Xauthority,display=:1'' ]
#
# To create one using the VNC backend and sensible defaults:
vfb = [ ''type=vnc'' ]
# The backend listens on 127.0.0.1 port 5900+N by default, where N is
# the domain ID. You can override both address and N:
#
# Or you can bind the first unused port above 5900:
# vfb = [ ''type=vnc,vnclisten=0.0.0.0,vnunused=1'' ]
# You can override the password:
# vfb = [ ''type=vnc,vncpasswd=MYPASSWD'' ]
# Set the hostname.
#hostname= "node%d" % vmid
# Set root device.
root = "/dev/hda1"
# Sets runlevel 2.
#extra = "2"
extra=''console=xvc0 xencons=xvc0''
#sdl=0
#vnc=0
nographics=0
--------------------------------------------------------------
node02:
--------------------------------------------------------------
# -*- mode: python; -*-
kernel="/boot/domU_64/vmlinuz-2.6.21-xxxx-xenU-ipv4-64"
memory = "256"
maxmem = "512"
name = "node02"
vcpus=1
#interface=eth0
dhcp="off"
vif = [ ''ip=10.0.0.2'' ] # replace with DomU''s
IP
ip = "10.0.0.2"
gateway= "10.0.0.128" # Dom0
netmask = "255.255.255.0"
# example for LVM or partition
disk = [ ''phy:/dev/mapper/xenvg-node02,/dev/hda1,w'',
''phy:/dev/mapper/xenvg-node02--swap,hda2,w'' ]
# example for image file
#disk = [ ''file:xen/xenvg-node0001.img,hda1,w'',
''file:/xen/xenvg-node0001-swap.img,hda2,w'' ]
# This uses environment variables XAUTHORITY and DISPLAY. You
# can override that:
# vfb = [
''type=sdl,xauthority=/home/bozo/.Xauthority,display=:1'' ]
#
# To create one using the VNC backend and sensible defaults:
vfb = [ ''type=vnc'' ]
# The backend listens on 127.0.0.1 port 5900+N by default, where N is
# the domain ID. You can override both address and N:
#
# Or you can bind the first unused port above 5900:
# vfb = [ ''type=vnc,vnclisten=0.0.0.0,vnunused=1'' ]
# You can override the password:
# vfb = [ ''type=vnc,vncpasswd=MYPASSWD'' ]
# Set the hostname.
#hostname= "node%d" % vmid
# Set root device.
root = "/dev/hda1"
# Sets runlevel 2.
#extra = "2"
extra=''console=xvc0 xencons=xvc0''
#sdl=0
#vnc=0
nographics=0
--------------------------------------------------------------
In node02 configuration file, if I specify interface=eth0, I get the
following error message: "Error: name ''eth0'' is not
defined". Also for
node02 I have tested with a 32bits Debian release and I had the same
problem.
Now here is the log file of a working VM:
--------------------------------------------------------------
Watching /local/domain/0/device-model/56/logdirty/next-active
Watching /local/domain/0/device-model/56/command
shift keysym 003e keycode 86
[...]
shift keysym 003f keycode 53
xs_read(): vncpasswd get error.
/vm/55ba1747-eb73-8321-da9c-44db8f906295-1/vncpasswd.
char device redirected to /dev/pts/1
Console: prepared domain, waiting for ringref at
/local/domain/56/console or /local/domain/56/serial/0
FB: Waiting for KBD backend creation
Doing backend watch on /local/domain/0/backend/vkbd/56/0
FB: Waiting for FB backend creation
Doing backend watch on /local/domain/0/backend/vfb/56/0
Console: got watch
Console: got ring-ref 149301 port 2
Console: connected to guest frontend
FB: Waiting for KBD frontend initialization
Doing frontend watch on /local/domain/56/device/vkbd/0
FB: Carry on waiting
--------------------------------------------------------------
And the log file of a VM that doesn''t start and uses 100% cpu in
xentop:
--------------------------------------------------------------
Watching /local/domain/0/device-model/57/logdirty/next-active
Watching /local/domain/0/device-model/57/command
shift keysym 003e keycode 86
[...]
shift keysym 003f keycode 53
xs_read(): vncpasswd get error.
/vm/a318301b-7221-8082-f302-4f2e4bec156b/vncpasswd.
bind() failed
--------------------------------------------------------------
And here is the ifconfig output on Dom0: (I have removed some data
that didn''t seem relevant)
--------------------------------------------------------------
# ifconfig
eth0 Link encap:Ethernet HWaddr 00:1B:FC:19:19:36
inet addr:91.121.77.156 Bcast:91.121.77.255 Mask:255.255.255.0
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
peth0 Link encap:Ethernet HWaddr 00:1B:FC:19:19:36
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
vif56.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF
inet addr:10.0.0.128 Bcast:0.0.0.0 Mask:255.255.255.255
vif57.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF
inet addr:10.0.0.129 Bcast:0.0.0.0 Mask:255.255.255.255
--------------------------------------------------------------
I have noticed that when only 1 VM is running I only have 1 vifxx.0
When I start a second VM (although it hangs) I see a second vif that
appears, is that normal?
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
On Sunday June 08 2008 08:32:47 am Thomas wrote:> @Jim: mac address in vif? I thought it was an IPv4 address I had to > put. What should I put then? Any figures?Xen recommends using a mac address in the block starting with 00:16:3e, so:> vif = [ ''ip=10.0.0.1'' ]becomes vif = [ ''mac=00:16:3e:jk:lm:no'' ]. Substitute unique numbers for jklmno in each different vm. You may keep the ip= parm, but it doesn''t seem to be doing what you want - see below.> disk = [ ''phy:/dev/mapper/xenvg-node01,/dev/hda1,w'', > ''phy:/dev/mapper/xenvg-node01--swap,hda2,w'' ]If you are using xen 3.1 or up, you can substitute xvda for each hda above for a small performance increase.> kernel="/boot/domU_32/vmlinuz-2.6.21-xxxx-xenU-ipv4-32"> kernel="/boot/domU_32/vmlinuz-2.6.21-xxxx-xenU-ipv4-64"Little confused: is one a 32 bit domain, and one a 64 bit domain? With xen 3.0.3, this may not work. Should work fine with xen 3.1 and up.> bind() failedHave no clue what that means.> vif56.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF > inet addr:10.0.0.128 Bcast:0.0.0.0 Mask:255.255.255.255 > > vif57.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF > inet addr:10.0.0.129 Bcast:0.0.0.0 Mask:255.255.255.255 > -------------------------------------------------------------- > > I have noticed that when only 1 VM is running I only have 1 vifxx.0 > When I start a second VM (although it hangs) I see a second vif that > appears, is that normal?Sounds normal. Part of domain creation is device creation. Once the 2nd domain is created tho'', it hangs. What concerns me is that one of your vifs has your gateway address. Not only that, vifs don''t normally have an ip at all in bridging mode. What''s the ifconfig output of your working vm? I don''t know how long the ip= parm in your config lasts during domain creation, but once the vm is up, it gets it''s own ip the same way any physical machine gets it - the way you set it up in system-config-network, or your distro''s equivalent. Make sure the vm has a static address set to 10.0.0.1 for the first vm, .02 for the second, etc. Good luck. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Jim, My Dom0 is Debian Etch 64bits. Therefore it is probably a good idea that I use 64bits DomU too just to lower the chances of running into troubles, I''ll do that right away. My Xen version is 3.2.1, well that''s what ovh (my web hosting provider) says. Strangely, when I do "xm info" I see: xen_caps : xen-3.0-x86_64 xen-3.0-x86_32p I cannot use bridging mode, therefore I use NAT, I should have mentioned that earlier, sorry about that. The ifconfig output for a working VM is as expected: eth0 ... 10.0.0.1, for the second VM I can''t know as it doesn''t work. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Sunday June 08 2008 11:02:50 am Thomas wrote:> My Dom0 is Debian Etch 64bits. Therefore it is probably a good idea > that I use 64bits DomU too just to lower the chances of running into > troubles, I''ll do that right away. > > My Xen version is 3.2.1, well that''s what ovh (my web hosting > provider) says. Strangely, when I do "xm info" I see: > xen_caps : xen-3.0-x86_64 xen-3.0-x86_32pYeah, caps never changed with 3.1 and up. The version you''ve been told should match something like: xen_major : 3 xen_minor : 2 xen_extra : .1 If it is indeed 3.2.1, 32-on-64 should work fine.> I cannot use bridging mode, therefore I use NAT, I should have > mentioned that earlier, sorry about that. > > The ifconfig output for a working VM is as expected: eth0 ... > 10.0.0.1, for the second VM I can''t know as it doesn''t work.That''s good. I guess NAT accounts for the vifs having ips, but then shouldn''t your 2nd vm''s gateway be 10.0.0.129, not .128? Your private email to Ray (that he cc-ed to the list) mentions your kernel release is 2.6.21.7. Curious - makes me think your Debian provider is using a Fedora kernel, as well as having updated to Xen 3.2.1. (Not relevant, just curious :-) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> Yeah, caps never changed with 3.1 and up. The version you''ve been told should > match something like: > > xen_major : 3 > xen_minor : 2 > xen_extra : .1 > > If it is indeed 3.2.1, 32-on-64 should work fine.Then yes my Xen is 3.2.1> That''s good. I guess NAT accounts for the vifs having ips, but then shouldn''t > your 2nd vm''s gateway be 10.0.0.129, not .128?Now that I specify the mac address in vif=[...], the vifxx.0 in Dom0 keeps changing. Therefore my gateway gets always wrong, and the network is no longer reachable. By the way, I have deactivated network support in xend-config.sxp, just to check, and still my second VM won''t run, so I think we can say that is nothing to do with networking. What could possibly go wrong? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Sunday June 08 2008 01:19:07 pm Thomas wrote:> > That''s good. I guess NAT accounts for the vifs having ips, but then > > shouldn''t your 2nd vm''s gateway be 10.0.0.129, not .128? > > Now that I specify the mac address in vif=[...], the vifxx.0 in Dom0 > keeps changing. Therefore my gateway gets always wrong, and the > network is no longer reachable.The ''xx'' part is the domain id #, and will change with each new xm create. Do you mean the ip for that vifxx.0 changes as well? If you can log in thru ssh (probably not w/o a gateway), vnc, or xm create -c, you can change your gateway in the domu with ''route'' till we get this straightened out. Anybody out there have a better idea how NAT works?> By the way, I have deactivated network support in xend-config.sxp, > just to check, and still my second VM won''t run, so I think we can say > that is nothing to do with networking. What could possibly go wrong?Umm, now I''m confused :-( Try posting your boot messages from as far as your 2nd vm gets. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ok, now I am able to add the mac address and have expected behavior in the Dom0 and DomU ifconfig output using: vif = [ ''mac=00:16:3e:00:00:01, ip=10.0.0.1'' ] with the proper gateway and netmask, the network is back in business using NAT and is reachable from anywhere on the planet. I still have the problem of booting a second VM. The problem is I can''t post any output from the boot message as I always get the following error message after issuing: ----- # xm create -c node02 ------ xenconsole: Could not read tty from store: No such file or directory ------ However I do see a new vif with IP 10.0.0.129 in Dom0, that''s what you were talking about previously. That''s all I have beside the log file I have posted earlier which is still the same: ------ xs_read(): vncpasswd get error. /vm/365db4c4-4efc-54dc-2b01-19287a470079/vncpasswd. bind() failed Gasp, I was almost there... I am sure it is some stupid configuration value to adjust that is messing everything up like this, and the documentation is so scarce... _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Sunday June 08 2008 02:53:02 pm Thomas wrote:> Ok, now I am able to add the mac address and have expected behavior in > the Dom0 and DomU ifconfig output using: > vif = [ ''mac=00:16:3e:00:00:01, ip=10.0.0.1'' ] > > with the proper gateway and netmask, the network is back in business > using NAT and is reachable from anywhere on the planet. > > I still have the problem of booting a second VM. The problem is I > can''t post any output from the boot message as I always get the > following error message after issuing: > ----- > # xm create -c node02 > ------ > xenconsole: Could not read tty from store: No such file or directory > ------ > However I do see a new vif with IP 10.0.0.129 in Dom0, that''s what you > were talking about previously.Nice! At least we haven''t lost anything! Unfortunately, I''ve never been able to get past that xenconsole error myself w/o a full reboot of dom0. I don''t know how easy that will be with your hosting service. Nor do I guarantee that that error will go away. I''ve only corrected what you could show me so far :-( _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Concerning the console problem I have found this: http://marc.info/?l=xen-users&m=118251303120172&w=2 ------------------ Hi Zechao, well I''ve had a similiar problem, maybe this will help you to solve the problem: Look for the following line in the file /etc/inittab of you domU (try to mount the domU image in dom0 to edit the inittab file of the vm): co:2345:respawn:/sbin/mingetty console If the line is missing simply add it to the inittab. Below you find an example how (a part of) my inittab looks like. ----- # Run gettys in standard runlevels # #line added to get a login prompt: #(dont forget to comment out the following line) # co:2345:respawn:/sbin/mingetty console #co:2345:respawn:/sbin/agetty xvc0 9600 vt100-nav #1:2345:respawn:/sbin/mingetty tty1 #2:2345:respawn:/sbin/mingetty tty2 #3:2345:respawn:/sbin/mingetty tty3 #4:2345:respawn:/sbin/mingetty tty4 #5:2345:respawn:/sbin/mingetty tty5 #6:2345:respawn:/sbin/mingetty tty6 ------------- Instead of co:2345:respawn:/sbin/mingetty console, I have put: co:2345:respawn:/sbin/gety console as I don''t have mingetty on my Debian. But I still get the same error message. I also found this: http://www.nabble.com/xenconsole:-Could-not-read-tty-from-store:-No-such-file-or-directory-td11245504.html But I''d like to know exactly what is going wrong, and if it applies to my Debian DomU? I don''t to mess things up even more then they already are. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Sunday June 08 2008 03:42:24 pm Thomas wrote:> co:2345:respawn:/sbin/mingetty console > #co:2345:respawn:/sbin/agetty xvc0 9600 vt100-nav[...]> Instead of co:2345:respawn:/sbin/mingetty console, > I have put: co:2345:respawn:/sbin/gety console as I don''t have > mingetty on my Debian.How about ''agetty''?> But I''d like to know exactly what is going wrong, and if it applies to > my Debian DomU? I don''t to mess things up even more then they already > are.No idea at this point. Sorry. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Argh! I noticed I made a typo, I typed gety instead of getty, but I still have the same problem. This is becoming very frustrating. Moreover the problem can come from my web hosting company that supplies the DomU distro ready to install which I am using. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
This morning I tried using a disk image instead of LVM for the second VM, and I still have the exact same problem. I am starting to become really suspicious about the DomU distro I am using. I will try to use another method for installing a DomU. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Now that you''ve posted your configs and the error you''re getting about vncpasswd, this is a little more clear. Comment out the line that says: vfb = [ ''type=vnc'' ] I realize that the default behavior (or at least the way I know it to be) is as your config mentions: # The backend listens on 127.0.0.1 port 5900+N by default, where N is # the domain ID. But clearly vnc can''t bind(), probably because it''s not incrementing the port number. I just duplicated this problem locally on Cent 5.1 with Xen 3.2.1 compiled from source. Try something more reasonable like: vfb = [ ''type=vnc,vnclisten=0.0.0.0,vncdisplay=2,vncpasswd=letsgetit'' ] This will give you a VNC instance on :5902, just modify vncdisplay for each domU. If I have time, I''ll dig into the code and find out if this is really a "bug". -Ray On 6/9/08, Thomas <iamkenzo@gmail.com> wrote:> This morning I tried using a disk image instead of LVM for the second > VM, and I still have the exact same problem. I am starting to become > really suspicious about the DomU distro I am using. I will try to use > another method for installing a DomU. > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ray!!!! You are fantastic, it works!!! I now have 2 VM up and running with access to the network using NAT!!!! I will name my kid Ray. Thank you all for your prompt assistance. Best regards, _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-08 14:33]:> Here are my config files:Because sometimes less is more, I used a stripped-down version of your configuration files:> node01:# -*- mode: python; -*- kernel="/boot/domU_64/vmlinuz-2.6.21-xxxx-xenU-ipv4-64" memory = "256" maxmem = "512" name = "node02" vcpus=1 vif = [ ''ip=10.0.0.2'' ] # replace with DomU''s IP ip = "10.0.0.2" gateway= "10.0.0.128" # Dom0 netmask = "255.255.255.0" disk = [ ''file:/xen/node01.img,hda1,w'', ''file:/xen/node01-swap.img,hda2,w'' ] root = "/dev/hda1" extra=''console=xvc0 xencons=xvc0'' nographics=0> -------------------------------------------------------------- > > node02: > --------------------------------------------------------------# -*- mode: python; -*- kernel="/boot/domU_32/vmlinuz-2.6.21-xxxx-xenU-ipv4-32" memory = "256" maxmem = "512" name = "node01" vcpus=1 vif = [ ''ip=10.0.0.1'' ] # replace with DomU''s IP ip = "10.0.0.1" gateway= "10.0.0.128" # Dom0 netmask = "255.255.255.0" disk = [ ''file:/xen/node02.img,hda1,w'', ''file:/xen/node02-swap.img,hda2,w'' ] root = "/dev/hda1" extra=''console=xvc0 xencons=xvc0'' nographics=0> --------------------------------------------------------------both boot up fine. I didn''t test networking yet, but If you''re going to use only NAT-based machines and none using your "failover-IP" I recommend changing /etc/xen/xend-config.sxp anyway (comment the route lines at lines 161, 162 and activate the NAT-lines at 167, 168)> I have noticed that when only 1 VM is running I only have 1 vifxx.0 > When I start a second VM (although it hangs) I see a second vif that > appears, is that normal?On the host, there is one vif per virtual interface. That is, if you have a VM with more than 1 NIC, you''ll see accordingly more vifs on the host. Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Felix Krohn <felix.krohn@ovh.net> [2008-06-09 14:15]:> [...]OK, too late, looks like ray solved it already :) Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Felix, You stripped down config files work perfectly also. Now I understand what is going on with Xen. Thank you. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-09 14:24]:> You stripped down config files work perfectly also. Now I understand > what is going on with Xen.Fine! :-) Don''t hesitate to contact me, If you have further problems or questions. Especially if there is something that may be caused by OVH''s Xen distribution, like it may be the case here. I will change the example configuration files to not include the vnc parameters by default, as they''re of limited use in that setup anyway (not to speak of VNC''s protocol weaknesses...) Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Mon, Jun 9, 2008 at 8:02 AM, Thomas <iamkenzo@gmail.com> wrote:> Ray!!!! You are fantastic, it works!!! I now have 2 VM up and running > with access to the network using NAT!!!! > > I will name my kid Ray.Not a bad choice :) I''ve reported this at http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1272 for anyone interested. -Ray _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hmmm, now for some reason I can''t "xm console" into my DomUs. However when I do "xm shutdown node01 && xm create -c node01", I do get the console, but when I do ctrl+], followed by "xm console node01", then nothing happens, hitting ctrl+c doesn''t help either. My VMs are still functional though. I added the vfb with the vnc parameters, but it didn''t help. Of course I could setup an iptables rule to forward ssh connections on some custom ports to each DomU, but I''d like to know what''s happening now with the "xm console". _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-10 00:00]:> console, but when I do ctrl+], followed by "xm console node01", then > nothing happens, hitting ctrl+c doesn''t help either. My VMs are stillYou also tried hitting just enter or "any key"? I believe you mentioned changing the /etc/inittab of the domU earlier on, you should check that you ot the line 0:2345:respawn:/sbin/getty 38400 xvc0 in it.> Of course I could setup an iptables rule to forward ssh connections on > some custom ports to each DomU, but I''d like to know what''s happening > now with the "xm console".Sounds like you''re knocking but nobody opens up :) Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Felix, I recreated my Dom0, and now it works. I must have edited so many config files that I might have forgotten to revert some to their initial state. Virtualization is fantastic, I can recreate full virtual servers in a matter of minutes, shut them down, bring them back to life, change their ram allocation... Thank you for providing this Debian Xen server. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ok I understand the problem. As soon as I set /etc/hostname in my DomU, I can''t xm console into it any more. Any thoughts on that? I wanted to set hostname, so that I don''t get mixed up and I know in which dom0 I am working. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-10 12:21]:> As soon as I set /etc/hostname in my DomU, I can''t xm console into it > any more. Any thoughts on that?Odd. I can''t reproduce it here, using the configuration file I posted yesterday. Are there any error messages? Does it work when you ''xm create -c'', and if so, until what point in the bootup process? Does it still show up in ''xm list'' or can you access it using ssh? Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
There is no error message, it just seem to hang. using xm create -c works and goes to the end of the boot process at the login prompt. The VM works perfectly and is displayed in xm list. I''ll leave it like this for now and put it on my todo list. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Argh it''s stuck again! It happens after some amount of time. I only have 1 DomU running, I can xm console into it after creating it, but if I wait 5-10 minutes, I can''t console into it anymore. The DomU is fully functionnal, appears in xm list and is not stuck at 100% cpu, and I can ssh into it. It is probably not related, but I noticed that the memory allocated to Dom0 gets lower and lower each time I add a DomU, and when the DomU is shutdown, it doesn''t get back its full memory. Currently my server has 2Gb ram, my Dom0 has 706Mb and my only DomU has 256Mb, I am missing 1Gb ram. As I can ssh into my DomU, it is not a big issue, but still I''d like to know what is going on, it''s good to be able to look at the console log if there are any future problems. Could it be related to vnc? because it seems there is a timing somewhere that blocks the xm console to DomU. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Mysteriously, I am now able to xm console in any DomU without any problem... I also have set /etc/hostname in all of them. Maybe after creating all my DomUs I need to wait a few minutes until I can console into any of them? Anyone has that same issue? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-10 14:42]:> Anyone has that same issue?I don''t know what''s going wrong. Can you see any errors in the domU logs or nothing at all? No /sbin/init related messages? Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Gone again, I can''t "xm console" anymore. I can still ssh though. I noticed that I had to allow INPUT from 10.0.0.0/16 in iptables in order to be able to ssh. Moreover ssh takes roughly 5s to prompt for my password. That''s strange as the DomU are on the same server, I would expect the connection to be a bit quicker. I then flushed all iptables rules hoping that I could console back in, but no luck. I still can''t xm console in any DomU. And there is nothing fancy in /var/log/xen Am I the only person to experience all these issues? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Karsten Nielsen
2008-Jun-10 13:21 UTC
Re: Fwd: [Xen-users] Unable to create more than 1 VM
try adding extra=''xencons=tty1'' to your domU config file and shutdown and start your domU - karsten Thomas wrote:> Gone again, I can''t "xm console" anymore. I can still ssh though. I > noticed that I had to allow INPUT from 10.0.0.0/16 in iptables in > order to be able to ssh. Moreover ssh takes roughly 5s to prompt for > my password. That''s strange as the DomU are on the same server, I > would expect the connection to be a bit quicker. > > I then flushed all iptables rules hoping that I could console back in, > but no luck. I still can''t xm console in any DomU. > > And there is nothing fancy in /var/log/xen > > Am I the only person to experience all these issues? > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-10 15:18]:> order to be able to ssh. Moreover ssh takes roughly 5s to prompt for > my password. That''s strange as the DomU are on the same server, I > would expect the connection to be a bit quicker.I guess your DNS is broken - make sure you can resolve names from inside the domU and that the adress you''re connecting from can be reversed (easiest thing is to add an entry in /etc/hosts in the domU)> And there is nothing fancy in /var/log/xen/var/log/syslog? Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> I guess your DNS is broken - make sure you can resolve names from inside > the domU and that the adress you''re connecting from can be reversed > (easiest thing is to add an entry in /etc/hosts in the domU)You are right, I can''t resolve names from inside a DomU. I tried doing "apt-get update" and it was unable to connect because it couldn''t resolve names. In the Dom0 it does work though. How do I fix that? I never run into such issue before. Do I need to install dnsutils? I cannot use nslookup in a DomU as it is not installed. Here is the /etc/resolv.conf in Dom0: ---------------------------------------------------- nameserver 127.0.0.1 nameserver my.server.IP.address nameserver 213.186.33.99 Here is the /etc/resolv.conf in a DomU: -------------------------------------------------------- nameserver 213.186.33.99 Is 213.186.33.99 the IP address of Ovh DNS server? Beside that when I ssh I use the IP address of the DomU, so I use the following command: "ssh user@10.0.0.1" therefore DNS is not part of the equation. Am I right?> /var/log/syslog?There is only stuff about cron, nothing special I could find.> I experienced the same thing. When I created a couple of guest, the > momory of Dom0 was reduced. However, when I deleted those guest, the > memory didn''t come back to its original state. Only after rebooting > Dom0, its memory went back to the number it should be. Maybe I''m > misunderstanding something, but that doesn''t seem a reasonable > behavior (I mean, having to restart Dom0). > GastonYeah really strange. It might be Xen''s default behaviour, there is probably a special command to give back the freed memory to Dom0. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-10 19:42]:> resolve names. In the Dom0 it does work though. How do I fix that? Itry using ''dig'' or ''host'', at least the latte is definitely installed.> Here is the /etc/resolv.conf in Dom0: > ---------------------------------------------------- > nameserver 127.0.0.1 > nameserver my.server.IP.addressdelete above 2 lines and keep the next one:> nameserver 213.186.33.99 > > Here is the /etc/resolv.conf in a DomU: > -------------------------------------------------------- > nameserver 213.186.33.99 > > Is 213.186.33.99 the IP address of Ovh DNS server?Yep. Keep that one.> Beside that when I ssh I use the IP address of the DomU, so I use the > following command: "ssh user@10.0.0.1" therefore DNS is not part of > the equation. Am I right?No, not entirely: sshd will make a reverse lookup ("what is the hostname of 10.x.x.x?") to put the host in the log and lets you wait until it is either resolved or the timeout occurs.> Yeah really strange. It might be Xen''s default behaviour, there is > probably a special command to give back the freed memory to Dom0.It''s still there - you maybe don''t see it, but you mem is still there and available. You can try ''xm mem-set'', but don''t need to. Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> try using ''dig'' or ''host'', at least the latte is definitely installed.Neither dig nor host are installed on my DomU-64 provided by Ovh. I did #updatedb, then #locate host, and there is nothing in /usr/bin. Dom0 however does have them installed by default.>> Is 213.186.33.99 the IP address of Ovh DNS server? > > Yep. Keep that one.Ok thanks.> No, not entirely: sshd will make a reverse lookup ("what is the hostname > of 10.x.x.x?") to put the host in the log and lets you wait until it is > either resolved or the timeout occurs.You were partially right. It is actually in DomU/etc/hosts that one should add the Dom0 address through which it connects to have blazing fast ssh login prompt.> It''s still there - you maybe don''t see it, but you mem is still there > and available. > You can try ''xm mem-set'', but don''t need to.Thank you Felix.> try adding extra=''xencons=tty1'' to your domU config file > > and shutdown and start your domU > > - karstenThanks Karsten, I already tried this option and we are almost there, I do get the console, but it stops at: [edited for brevity] VFS: Mounted root (ext3 filesystem). Freeing unused kernel memory: 240k freed Adding 262136k swap on /dev/hda2. Priority:-1 extents:1 across:262136k EXT3 FS on hda1, internal journal [End] Then I don''t see any login prompt. There might again an issue in the config file, it must be some parameter to add to sort things out. The DomU is fully functionnal though, and I can still ssh into it. Can someone provide me a working config file, so that I can pick bits and pieces from it and do a lot of trial and error to nail the problem down? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
The "xm console" is driving me nuts. I use the following option in the DomU config file : extra=''console=xvc0 xencons=xvc0'' I can xm console right after creating a DomU, but if I wait too long, it won''t work anymore. Is there a timeout somewhere that is blocking me? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Concerning the name resolution here is my problem:>From Dom0, I can ping the whole planet.>From DomU, I can ping any address in 10.0.0.0, I can ping my serverpublic address 91.xxx.xxx.xxx, but I cannot ping any other IP address. Even Ovh''s DNS server 213.186.33.99 cannot be pinged. I have flushed any iptables rule except the following NAT redirection: -A PREROUTING -p tcp -i eth0 --dport 80 -j DNAT --to 10.0.0.1:80 This is for connections from the outside world that come into the server. Do I need to add another rule for connections from a DomU to the outside world? I guess I need to. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Concerning Dom0 not getting its memory back, please read this article: http://searchservervirtualization.techtarget.com/tip/0,289483,sid94_gci1284833,00.html Excerpt: ----------------- When a virtual machine starts, it normally takes the memory that it has assigned from the memory that is available to Dom0. Once allocated, the Dom0 will never get that memory back, not even when the virtual machines are all stopped. It is dated from 12.04.2007, probably Xen changed its behavior since. Can someone confirm? The article also talks about CPU allocation which can be very handy if you use multi-core CPUs. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Tuesday June 10 2008 07:40:59 am Thomas wrote:> Argh it''s stuck again! It happens after some amount of time. I only > have 1 DomU running, I can xm console into it after creating it, but > if I wait 5-10 minutes, I can''t console into it anymore. The DomU is > fully functionnal, appears in xm list and is not stuck at 100% cpu, > and I can ssh into it.Hi Again Thomas. This sounds like a symptom of the getty dying on xvc0. Are you getting messages of the form "co:init getty restarting too fast. delaying for 5 min." or something to that effect? The syslog is one of /var/log/{syslog,messages}.> > try adding extra=''xencons=tty1'' to your domU config file > > > > and shutdown and start your domU > > > > - karsten > > Thanks Karsten, I already tried this option and we are almost there, I > do get the console, but it stops at: > > [edited for brevity] > VFS: Mounted root (ext3 filesystem). > Freeing unused kernel memory: 240k freed > Adding 262136k swap on /dev/hda2. Priority:-1 extents:1 across:262136k > EXT3 FS on hda1, internal journal > [End] > > Then I don''t see any login prompt. There might again an issue in the > config file, it must be some parameter to add to sort things out. The > DomU is fully functionnal though, and I can still ssh into it.This is normal end of boot process messages, but you are not getting a login prompt. I remember you had problems adding "co:2345:respawn:/sbin/agetty xvc0 9600 vt100-nav" to your inittab because you didn''t have ''agetty'' (which is in the util-linux package (SuSE) or util-linux-ng (Fedora), usually in /sbin). If you can''t find it, try one of the other getty programs in /sbin or /usr/sbin. Just as important is to put xvc0 in your /etc/securetty.> Excerpt: > ----------------- > When a virtual machine starts, it normally takes the memory that it > has assigned from the memory that is available to Dom0. Once > allocated, the Dom0 will never get that memory back, not even when the > virtual machines are all stopped. > > It is dated from 12.04.2007, probably Xen changed its behavior since. > Can someone confirm?Nope, still the case, at least if you mean automatically gets it back. You have to manually add it back with the ''xm mem-set 0 mem-size'' command. It makes me nervous to give dom0 more memory than it started out with :-), so make sure you remember what ''xm list'' said for Domain-0 before you started any domus. It''s not the same as the amount of memory installed, or even what ''xm info'' reports. So for each domu that exits, you can add that domu''s amount of memory to the current amount in ''xm list 0'', So for example, on my system, even tho'' xm info reports: total_memory : 2038 my initial dom0 memory is about 1957. So if I have [535] > xm list Name ID Mem VCPUs State Time(s) Domain-0 0 956 2 r----- 192036.6 fedora 2 512 1 -b---- 22698.5 winxp 29 512 2 -b---- 23504.2 and kill those two domus, even tho'' 956 + 512 + 512 is greater than 1957, I wouldn''t go higher than 1957 with xm mem-set 0 ... This really isn''t something I worry about unless I start getting error messages about not having enough memory to start a domu. Xen *usually* is good about reusing freed memory for domus, just not giving it back to dom0. Also, setting dom0-min-mem in /etc/xend-config.sxp might give you some peace of mind that dom0 will always have what it needs. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Jim, while I was checking out /var/log/syslog/* I couldn''t find anything special about getty restarting too fast. I will give a try at your ideas. xvc0 is already present in /etc/securetty in DomU. And thank you for your clarification about the memory management in Dom0. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> Hi Again Thomas. This sounds like a symptom of the getty dying on xvc0. Are > you getting messages of the form "co:init getty restarting too fast. delaying > for 5 min." or something to that effect? The syslog is one > of /var/log/{syslog,messages}.Hi Jim, I am back, and I have good news, I left my console opened after it was created, and look what I have found after 5 minutes: ------- INIT: Id "1" respawning too fast: disabled for 5 minutes INIT: no more processes left in this runlevel ------- Here is the configuration file for my DomU: ------ # -*- mode: python; -*- kernel = "/boot/vmlinuz-2.6.18-4-xen-amd64" ramdisk = "/boot/initrd.img-2.6.18-4-xen-amd64" memory = "256" maxmem = "512" name = "VM1" vcpus=1 vif = [ ''mac=00:16:3e:00:00:01, ip=10.0.0.1'' ] # example for LVM or partition disk = [ ''phy:/dev/mapper/xenvg-VM1,/dev/hda1,w'', ''phy:/dev/mapper/xenvg-VM1--swap,hda2,w'' ] root = "/dev/hda1" extra=''console=xvc0 xencons=xvc0'' nographics=0 ------- In the /etc/securetty of the DomU I have the following lines: ----- console # for xen: xvc0 ----- And in the /etc/inittab of my DomU: ----- 1:2345:respawn:/sbin/getty 38400 xvc0 ----- I was previously set to 0:2345:respawn...., and give the same error message. We are getting closer to the key! _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
I have to add the error message showed up when switching from a previous custom 2.6.21 kernel provided by ovh to the 2.6.18 official xen kernel. Acutally now I never see the login prompt at all. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
In the /etc/inittab of my DomU, I have put the following line and commented out the other similar one: ---- 1:2345:respawn:/sbin/getty 38400 tty1 ---- Now, I am able to get the login prompt when doing "xm create -c VM1", but if I detach from the console, I can''t console back in it. So I changed it to: ---- 1:2345:respawn:/sbin/getty 38400 xvc0 ---- And now when I do "xm create -c VM1", then I don''t get the login prompt anymore. I feel like I am running in circles... Am I really the only one having this problem? I will try installing a fresh pristine Debian Etch VM with debootstrap. I was currently using DomU images provided by Ovh. On 6/16/08, Thomas <iamkenzo@gmail.com> wrote:> I have to add the error message showed up when switching from a > previous custom 2.6.21 kernel provided by ovh to the 2.6.18 official > xen kernel. Acutally now I never see the login prompt at all. >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
In my DomU, when I login through ssh, I figured out that I don''t have any /dev/xvc0, could that be a symptom of my problem? On 6/16/08, Thomas <iamkenzo@gmail.com> wrote:> In the /etc/inittab of my DomU, I have put the following line and > commented out the other similar one: > ---- > 1:2345:respawn:/sbin/getty 38400 tty1 > ---- > > Now, I am able to get the login prompt when doing "xm create -c VM1", > but if I detach from the console, I can''t console back in it. > > So I changed it to: > ---- > 1:2345:respawn:/sbin/getty 38400 xvc0 > ---- > > And now when I do "xm create -c VM1", then I don''t get the login prompt > anymore. > > I feel like I am running in circles... Am I really the only one having > this problem? > > I will try installing a fresh pristine Debian Etch VM with > debootstrap. I was currently using DomU images provided by Ovh. > > On 6/16/08, Thomas <iamkenzo@gmail.com> wrote: >> I have to add the error message showed up when switching from a >> previous custom 2.6.21 kernel provided by ovh to the 2.6.18 official >> xen kernel. Acutally now I never see the login prompt at all. >> >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
* Thomas <iamkenzo@gmail.com> [2008-06-16 23:06]:> In my DomU, when I login through ssh, I figured out that I don''t have > any /dev/xvc0, could that be a symptom of my problem?Yes. try this inside domU: mknod /dev/xvc0 c 204 191 Amicalement, Félix -- Félix Krohn, <felix.krohn@ovh.net> _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Monday June 16 2008 02:37:20 pm Thomas wrote:> Hi Jim, > > I am back, and I have good news, I left my console opened after it was > created, and look what I have found after 5 minutes: > ------- > INIT: Id "1" respawning too fast: disabled for 5 minutes > INIT: no more processes left in this runlevel > ------- > > Here is the configuration file for my DomU:Hi, Thomas.> extra=''console=xvc0 xencons=xvc0''My extra= line says: extra=''console=xvc0 xencons=tty''> And in the /etc/inittab of my DomU: > ----- > 1:2345:respawn:/sbin/getty 38400 xvc0The ''1'' part of that line is telling getty to start up on tty1. ''2'' would start up on tty2, etc. When you tell it to use /dev/xvc0 for tty1, it fails, and you can''t login. When you change xvc0 back to tty1, you can login. The standard xen line for inittab uses ''co'', and is in addition to the other getty lines: co:2345:respawn:/sbin/getty xvc0 9600 vt100-nav Don''t know if the order of ''xvc0'' and ''9600'' is significant.> In my DomU, when I login through ssh, I figured out that I don''t have > any /dev/xvc0, could that be a symptom of my problem?Definitely. Was that under 2.6.18 or 2.6.21? Did Felix''s suggestion help? The standard xen 2.6.18 *should* have created xvc0 for you. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
It still doesn''t work. And my kernel is the one from the Debian package for Xen 2.6.18. I still don''t have any /dev/xvc0. I do have a /dev/xen/evtchn though. Does it help? Now what''s interesting is that when I do "xm create -c VM1", I can login to my DomU, and if I let it hang there for a while, while I am logged in, I still get the error message printed on my screen, I found that out while editing a file in vim, and suddenly after a few minutes of inactivity: INIT: Id "co" respawning too fast: disabled for 5 minutes. However I still have control over the console. We are just so close of the solution... So here is my DomU configuration file: ---- # -*- mode: python; -*- kernel = "/boot/vmlinuz-2.6.18-4-xen-amd64" ramdisk = "/boot/initrd.img-2.6.18-4-xen-amd64" memory = "256" name = "VM1" vif = [ ''mac=00:16:3e:00:00:01, ip=10.0.0.1'' ] disk = [ ''phy:/dev/mapper/xenvg-VM1,/dev/hda1,w'', ''phy:/dev/mapper/xenvg-VM1--swap,hda2,w'' ] root = "/dev/hda1" extra = "console=xvc0 xencons=tty" ---- And the inittab of my DomU (I have removed some commented lines for brevity): ---- # /etc/inittab: init(8) configuration. # $Id: inittab,v 1.91 2002/01/25 13:35:21 miquels Exp $ # The default runlevel. id:2:initdefault: # Boot-time system configuration/initialization script. # This is run first except when booting in emergency (-b) mode. si::sysinit:/etc/init.d/rcS # What to do in single-user mode. ~~:S:wait:/sbin/sulogin # /etc/init.d executes the S and K scripts upon change # of runlevel. # # Runlevel 0 is halt. # Runlevel 1 is single-user. # Runlevels 2-5 are multi-user. # Runlevel 6 is reboot. l0:0:wait:/etc/init.d/rc 0 l1:1:wait:/etc/init.d/rc 1 l2:2:wait:/etc/init.d/rc 2 l3:3:wait:/etc/init.d/rc 3 l4:4:wait:/etc/init.d/rc 4 l5:5:wait:/etc/init.d/rc 5 l6:6:wait:/etc/init.d/rc 6 # Normally not reached, but fallthrough in case of emergency. z6:6:respawn:/sbin/sulogin # What to do when CTRL-ALT-DEL is pressed. ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now # What to do when the power fails/returns. pf::powerwait:/etc/init.d/powerfail start pn::powerfailnow:/etc/init.d/powerfail now po::powerokwait:/etc/init.d/powerfail stop co:2345:respawn:/sbin/getty xvc0 9600 vt100-nav 1:2345:respawn:/sbin/getty 38400 tty1 ---- And here is my /etc/securetty of DomU: ---- # /etc/securetty: list of terminals on which root is allowed to login. # See securetty(5) and login(1). console # for xen: xvc0 # for people with serial port consoles ttyS0 # for devfs tts/0 # Standard consoles tty1 tty2 [...] ---- _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Tuesday June 17 2008 06:22:17 am Thomas wrote:> t still doesn''t work. And my kernel is the one from the Debian > package for Xen 2.6.18. I still don''t have any /dev/xvc0. I do have a > /dev/xen/evtchn though. Does it help?Not important to your problem.> Now what''s interesting is that when I do "xm create -c VM1", I can > login to my DomU, and if I let it hang there for a while, while I am > logged in, I still get the error message printed on my screen, I found > that out while editing a file in vim, and suddenly after a few minutes > of inactivity: > INIT: Id "co" respawning too fast: disabled for 5 minutes.If you look at a listing from ''/bin/ps a -HA'' (threaded child process display), you will see that getty is no longer in control of your login console. ''login'' is, with a child process for your login shell. So long as you don''t log out, ''co'' can respawn to it''s heart''s content :-) It''s after you log out, that the getty can''t be reestablished, and so control can''t be passed back to ''login''. (The fact that you can login at all tells me that /dev/xvc0 does exist, at least for the initial login. What happens after that is a mystery.) My guess is since you can''t find ''agetty'', that ''getty'' is a less stable substitute. (That, and the instability of xvc0 itself.) I''m going to chalk this up to problems with 2.6.18, and you may have to live with it. (That is, don''t log out after the initial login in ''xm create -c''.) With ''yum'', you can do slick things like say ''yum install /sbin/agetty'', and it will find the package that contains that program. Don''t know if there are any ''apt'' tools that can do the same.> # The default runlevel. > id:2:initdefault:This is telling Debian to boot into multiuser mode w/o nfs or X11. You might want to change that ''2'' to ''5'', tho'' I''m sure that Debian is doing an ''init 5'' somewhere along the way if you are getting vnc access now. (Just an interesting aside.) _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Once again thank you for your precious help. I will look at all this new info tomorrow, and hopefully come back with good news. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hmmm while creating a new Debian Etch VM using debootstrap, I found this in the boot messages of my VM: --- Xen virtual console successfully installed as tty1 --- _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users