[opensuse] installing leap42 as a xen host
jdd, I though I'd write up what I do when I install a xen host, just in case you spot a difference. Apart from installing over pxe+ssh+http, it's pretty standard stuff : boot leap42 installation, from local copy of repo. ssh login start yast select additional repos (leap updates) partitioning - (10G for Dom0, 1Gb swap, the rest for LVM). dom0 root filesystem, either ext4 or jfs. minimum server selection. software selection: add kernel-xen, remove kernel-default, add xen, add xen-tools. That's all I do. Interestingly, I think I have just reproduced your situation - on a Fujitsu desktop machine, I left the installation to finish by itself at around 16:15. When I got back around 17, it was in some weird state, but responding. I think I hit enter a couple of times and then it looked like I had a working system, just very slow. I've just now rebooted it, and it's now also just blank screen. I'm going to wait it out and see what happens - next thing is to fiddle with the grub options and get some more output. -- Per Jessen, Zürich (4.8°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Per Jessen wrote:
jdd,
I though I'd write up what I do when I install a xen host, just in case you spot a difference. Apart from installing over pxe+ssh+http, it's pretty standard stuff :
boot leap42 installation, from local copy of repo. ssh login start yast select additional repos (leap updates) partitioning - (10G for Dom0, 1Gb swap, the rest for LVM). dom0 root filesystem, either ext4 or jfs. minimum server selection. software selection: add kernel-xen, remove kernel-default, add xen, add xen-tools.
That's all I do. Interestingly, I think I have just reproduced your situation - on a Fujitsu desktop machine, I left the installation to finish by itself at around 16:15. When I got back around 17, it was in some weird state, but responding. I think I hit enter a couple of times and then it looked like I had a working system, just very slow. I've just now rebooted it, and it's now also just blank screen. I'm going to wait it out and see what happens - next thing is to fiddle with the grub options and get some more output.
On my machine, I saw numerous stalls: rcu_sched self-detected stall on cpu ... For the first boot-up, be patient, maybe wait 10minutes, then go and hit Esc and you should see some output. I pressed enter a few times which got the process moving and I finally got a login-prompt. The machine was still sluggish though. I amended the grub cmdline - I think it said "noresume resume=/dev/sda2 splash=silent ...." I changed this to "noresume". I don't know where to put "dom0_mem=512M". When you boot with that and you see the machine halting, just press Enter a couple of times. Something is clearly wrong, but at least you get a system up and running. -- Per Jessen, Zürich (3.6°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Per Jessen wrote:
Per Jessen wrote:
jdd,
I though I'd write up what I do when I install a xen host, just in case you spot a difference. Apart from installing over pxe+ssh+http, it's pretty standard stuff :
boot leap42 installation, from local copy of repo. ssh login start yast select additional repos (leap updates) partitioning - (10G for Dom0, 1Gb swap, the rest for LVM). dom0 root filesystem, either ext4 or jfs. minimum server selection. software selection: add kernel-xen, remove kernel-default, add xen, add xen-tools.
That's all I do. Interestingly, I think I have just reproduced your situation - on a Fujitsu desktop machine, I left the installation to finish by itself at around 16:15. When I got back around 17, it was in some weird state, but responding. I think I hit enter a couple of times and then it looked like I had a working system, just very slow. I've just now rebooted it, and it's now also just blank screen. I'm going to wait it out and see what happens - next thing is to fiddle with the grub options and get some more output.
On my machine, I saw numerous stalls:
rcu_sched self-detected stall on cpu ...
For the first boot-up, be patient, maybe wait 10minutes, then go and hit Esc and you should see some output. I pressed enter a few times which got the process moving and I finally got a login-prompt. The machine was still sluggish though. I amended the grub cmdline - I think it said
"noresume resume=/dev/sda2 splash=silent ...."
I changed this to "noresume". I don't know where to put "dom0_mem=512M".
When you boot with that and you see the machine halting, just press Enter a couple of times. Something is clearly wrong, but at least you get a system up and running.
I see some odd messages: guest57:~ # grep mod.*found /var/log/messages Feb 12 16:15:10 guest57 env[346]: modprobe: FATAL: Module xen-evtchn not found. Feb 12 16:15:10 guest57 env[355]: modprobe: FATAL: Module xen-gntdev not found. Feb 12 16:15:10 guest57 env[357]: modprobe: FATAL: Module xen-gntalloc not found. Feb 12 16:15:10 guest57 env[359]: modprobe: FATAL: Module xen-blkback not found. Feb 12 16:15:10 guest57 env[361]: modprobe: FATAL: Module xen-netback not found. Feb 12 16:15:10 guest57 env[363]: modprobe: FATAL: Module xen-pciback not found. Feb 12 16:15:11 guest57 modprobe[491]: modprobe: FATAL: Module xen-acpi-processor not found. Feb 12 17:27:27 linux-t6su modprobe[342]: modprobe: FATAL: Module xen-evtchn not found. Feb 12 17:27:27 linux-t6su modprobe[348]: modprobe: FATAL: Module xen-gntdev not found. Feb 12 17:27:27 linux-t6su modprobe[350]: modprobe: FATAL: Module xen-gntalloc not found. Feb 12 17:27:27 linux-t6su modprobe[352]: modprobe: FATAL: Module xen-blkback not found. Feb 12 17:27:27 linux-t6su modprobe[354]: modprobe: FATAL: Module xen-netback not found. Feb 12 17:27:27 linux-t6su modprobe[356]: modprobe: FATAL: Module xen-pciback not found. Feb 12 17:32:32 linux-t6su modprobe[517]: modprobe: FATAL: Module xen-acpi-processor not found. Feb 12 17:39:06 linux-t6su modprobe[341]: modprobe: FATAL: Module xen-evtchn not found. Feb 12 17:39:06 linux-t6su env[353]: modprobe: FATAL: Module xen-gntdev not found. Feb 12 17:39:06 linux-t6su modprobe[359]: modprobe: FATAL: Module xen-gntalloc not found. Feb 12 17:39:06 linux-t6su modprobe[361]: modprobe: FATAL: Module xen-blkback not found. Feb 12 17:39:06 linux-t6su modprobe[363]: modprobe: FATAL: Module xen-netback not found. Feb 12 17:39:07 linux-t6su modprobe[367]: modprobe: FATAL: Module xen-pciback not found. Feb 12 17:39:07 linux-t6su modprobe[556]: modprobe: FATAL: Module xen-acpi-processor not found. -- Per Jessen, Zürich (3.4°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/02/2016 18:06, Per Jessen a écrit :
For the first boot-up, be patient, maybe wait 10minutes, then go and hit Esc and you should see some output. I pressed enter a few times which got the process moving and I finally got a login-prompt. The machine was still sluggish though. I amended the grub cmdline - I think it said
"noresume resume=/dev/sda2 splash=silent ...."
I changed this to "noresume". I don't know where to put "dom0_mem=512M".
I added it to the two possible lines (multiboot and module) after removing all the multiboot options, I could see the boot go up to initramfs
When you boot with that and you see the machine halting, just press Enter a couple of times. Something is clearly wrong, but at least you get a system up and running.
not for me till now. and I would love to have grub accept to repeat key press :-( jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
jdd wrote:
Le 12/02/2016 18:06, Per Jessen a écrit :
For the first boot-up, be patient, maybe wait 10minutes, then go and hit Esc and you should see some output. I pressed enter a few times which got the process moving and I finally got a login-prompt. The machine was still sluggish though. I amended the grub cmdline - I think it said
"noresume resume=/dev/sda2 splash=silent ...."
I changed this to "noresume". I don't know where to put "dom0_mem=512M".
I added it to the two possible lines (multiboot and module) after removing all the multiboot options, I could see the boot go up to initramfs
When you boot with that and you see the machine halting, just press Enter a couple of times. Something is clearly wrong, but at least you get a system up and running.
not for me till now.
Did you wait 10-15minutes after the first blank screen turns up? Once I got the cmdline changed to "noresume", it was easy tell when it stalled again and then get it going by pressing Enter. I now have ssh access and the system seems to be working normally - I don't know if that is because the network interface keeps generating interrupts. -- Per Jessen, Zürich (3.2°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Per Jessen wrote:
jdd wrote:
Le 12/02/2016 18:06, Per Jessen a écrit :
For the first boot-up, be patient, maybe wait 10minutes, then go and hit Esc and you should see some output. I pressed enter a few times which got the process moving and I finally got a login-prompt. The machine was still sluggish though. I amended the grub cmdline - I think it said
"noresume resume=/dev/sda2 splash=silent ...."
I changed this to "noresume". I don't know where to put "dom0_mem=512M".
I added it to the two possible lines (multiboot and module) after removing all the multiboot options, I could see the boot go up to initramfs
When you boot with that and you see the machine halting, just press Enter a couple of times. Something is clearly wrong, but at least you get a system up and running.
not for me till now.
Did you wait 10-15minutes after the first blank screen turns up? Once I got the cmdline changed to "noresume", it was easy tell when it stalled again and then get it going by pressing Enter.
My system is in the datacentre downstairs, so I'm not always next to it to press enter every now and then - which causes a reboot to take forever. I don't know what these stalls and starvations are about, I'm waiting for Andrei to enlighten me. -- Per Jessen, Zürich (3.1°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/02/2016 18:46, Per Jessen a écrit :
Did you wait 10-15minutes after the first blank screen turns up?
even much more Once I
got the cmdline changed to "noresume", it was easy tell when it stalled again and then get it going by pressing Enter.
it goes until uncompressing initrd but I just noticed I have vmlinuz-xen and initrd-xen 4.1 but xen- 4.2 odd? I will control this, but now I have to go to kitchen :-)) jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
jdd wrote:
Le 12/02/2016 18:46, Per Jessen a écrit :
Did you wait 10-15minutes after the first blank screen turns up?
even much more
Once I
got the cmdline changed to "noresume", it was easy tell when it stalled again and then get it going by pressing Enter.
it goes until uncompressing initrd
but I just noticed I have vmlinuz-xen and initrd-xen 4.1 but xen- 4.2
odd?
I have kernel 4.1.15-8-xen, xen 4.5.2. The xen hypervisor version is independent of the kernel version, and the initrd file can be named anything, although the convention is that initrd- <version> matches vmlinuz-<version>. The stalling continues on my system, on start-up and shutdown both. It's essentially unusable. To reboot I have to manually go and get it going by pressing enter now and again. /Per -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 12/02/2016 21:02, Per Jessen a écrit :
The stalling continues on my system, on start-up and shutdown both. It's essentially unusable. To reboot I have to manually go and get it going by pressing enter now and again.
I completed bugzilla, maybe you can add yourself? I noticed that the boot went more with noremote and videotaped it. Seeing "journal started", I make the following: start a normal session, write down the time, wait some minutes, reboot on xen, wait some minutes, stop xen. wait... reboot a normal session. all this to make clear at what time what was started. It's pretty simple, as journalctl writes -- reboot --- behing two reboots... and no journal for xen... no change with "3" nor vga= on the command line despite what the image seems to say. a bit short of idea... my keyboard do not respond. I at install, specified to launch sshd and open the FW, but no ssh answer nor ping jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, 2016-02-12 at 18:46 +0100, Per Jessen wrote:
Did you wait 10-15minutes after the first blank screen turns up? Once I got the cmdline changed to "noresume", it was easy tell when it stalled again and then get it going by pressing Enter. I now have ssh access and the system seems to be working normally - I don't know if that is because the network interface keeps generating interrupts.
Years ago I had problems with a machine that would simply stop. The screen would freeze, the clock would stop moving. But move the mouse, hit a key, and the clock would jump the lost seconds (or minutes) and everything normal. It happened after a kernel update, and we never found the cause in bugzilla. The solution was to permanently have another machine ping this one every few seconds, to awake it. Ethernet activity causes interrupts, and this broke the spell. I still have the hardware, though. - -- Cheers, Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iEYEARECAAYFAla/P/8ACgkQtTMYHG2NR9U2dACeMYVxyrYHPwN3Gx151cERMUHq ToQAn2PsM0Nw6TL+KeyOBhzsA2JIligH =yuj9 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
solved details is yet to be found. I wanted to add some video memory, but it was not possible, but I could find some questionable options. I swapped three option and after that all works as expected. I can't know which because I have an install running on xen :-) Dell Optiplex 755 bios options swapped: Speedstep (was on set to off) VT was on set to off, but I hope it's not this one, very useful :-) CPUID limit ws off, set to on (I'm not sure) jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 13/02/2016 15:59, jdd a écrit :
Dell Optiplex 755
bios options swapped:
Speedstep (was on set to off) VT was on set to off, but I hope it's not this one, very useful :-) CPUID limit ws off, set to on (I'm not sure)
jdd
bad news. The problem is VT-X when on xen do not start. When off, no problem (but speed!) jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Le 13/02/2016 15:59, jdd a écrit : bad news. The problem is VT-X when on xen do not start. When off, no problem (but speed!) well... more precisely, it's "VT for direct I/O" and many xen modules do not start like per said details on the bug report https://bugzilla.opensuse.org/show_bug.cgi?id=966449 and (may be) http://dodin.info/wiki/pmwiki.php?n=Doc.Xen jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (3)
-
Carlos E. R. -
jdd -
Per Jessen