Hello all Xen devels, I''m writing howto''s in brazilian portuguese language about Xen-3.3.0 stubdom, pv-grub and others configurations. Following what is written in xen-3.3.0/stubdom/README, I tried start a HVM but without success. When I start the virtual machine calling ''xm create hvmtest'' command, in a few seconds the virtual machine is dead (see qemu-dm-hvmtest.log). This is happening following two distincts vnc display configurations: using a VNC server in the stub domain and using a VNC server in dom0. Any idea? Best regards, -- Marco Sinhoreli _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Fri 29 Aug 2008 12:49:13 -0300, a écrit :>Failed allocation for dom 30: 2048 extents of order 0>Failed to populate physmap>Failed set_mm_mapping>NULL vram_pointer See the first paragraph of the General Configuration section: there is a ballooning race condition which was discovered too late to get fixed in 3.3, so you need to restrict dom0''s memory usage thanks to the dom0_mem option. Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hello Samuel, I haven't seen the restrict memory to dom0, sorry! I set it in menu.lst but I have now another issue decribbed as follows: ******************* BLKFRONT for /local/domain/2/device/vbd/5632 ********** backend at /local/domain/0/backend/vbd/2/5632 Failed to read /local/domain/0/backend/vbd/2/5632/feature-flush-cache. 1299724 sectors of 512 bytes ************************** blk_open(/local/domain/2/device/vbd/5632) -> 5 xs_read(/local/domain/2/device/vbd/768/phantom_vbd): ENOENT ******************* BLKFRONT for /local/domain/2/device/vbd/768 ********** backend at /local/domain/0/backend/vbd/2/768 Failed to read /local/domain/0/backend/vbd/2/768/feature-flush-cache. 20971520 sectors of 512 bytes ************************** blk_open(/local/domain/2/device/vbd/768) -> 6 xs_directory(/local/domain/2/device/vkbd): ENOENT xs_directory(/local/domain/2/device/vfb): ENOENT xs_watch(/local/domain/0/device-model/1/logdirty/next-active, logdirty) Watching /local/domain/0/device-model/1/logdirty/next-active xs_watch(/local/domain/0/device-model/1/command, dm-command) Watching /local/domain/0/device-model/1/command xs_read(/local/domain/0/backend/pci/2/0/num_devs): EACCES ************************ NETFRONT for device/vif/1 ********** net TX ring size 256 net RX ring size 256 backend at /local/domain/0/backend/vif/2/1 mac is 00:16:3e:39:c7:b4 ************************** tap_open(device/vif/1) -> 7 open(/usr/share/xen/qemu/keymaps/pt-br, 0)-> 8 open(/usr/share/xen/qemu/keymaps/common, 0)-> 9 open(/usr/share/xen/qemu/keymaps/modifiers, 0)-> 10 close(10) close(9) close(8) sxs_read_watch() -> /local/domain/0/device-model/1/command dm-command xs_read(/local/domain/0/device-model/1/command): ENOENT xs_read_watch() -> /local/domain/0/device-model/1/logdirty/next-active logdirty xs_read(/local/domain/0/device-model/1/logdirty/key): ENOENT xs_read_watch() -> /local/domain/0/backend/vbd/2/5632/params hdc I commented only "vfb = [ 'type=vnc' ]" in hvmtest-dm file. Best regards, -- Marco Sinhoreli On Fri, Aug 29, 2008 at 12:56 PM, Samuel Thibault <samuel.thibault@ens-lyon.org> wrote:> Marco Sinhoreli, le Fri 29 Aug 2008 12:49:13 -0300, a écrit : >> > Failed allocation for dom 30: 2048 extents of order 0 >> > Failed to populate physmap >> > Failed set_mm_mapping >> > NULL vram_pointer > > See the first paragraph of the General Configuration section: there is a > ballooning race condition which was discovered too late to get fixed in > 3.3, so you need to restrict dom0's memory usage thanks to the dom0_mem > option. > > Samuel >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Fri 29 Aug 2008 13:35:38 -0300, a écrit :> Failed to read /local/domain/0/backend/vbd/2/5632/feature-flush-cache.(That''s benign)> xs_directory(/local/domain/2/device/vkbd): ENOENT > xs_directory(/local/domain/2/device/vfb): ENOENTNot a problem either, since your configuration makes ioemu use a VNC server.> xs_read(/local/domain/0/backend/pci/2/0/num_devs): EACCESNot a problem since you are not passing through anything.> sxs_read_watch() -> /local/domain/0/device-model/1/command dm-command > xs_read(/local/domain/0/device-model/1/command): ENOENT > xs_read_watch() -> /local/domain/0/device-model/1/logdirty/next-active logdirty > xs_read(/local/domain/0/device-model/1/logdirty/key): ENOENTThese are fine as well. So, what''s the problem? :) Does the hvmtest-dm domain die? Is the guest running? Is your network configuration OK to access the VNC server? (it may be easier to just start by disabling VNC and use a mere PVFB). Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hi Samuel,> So, what''s the problem? :) > Does the hvmtest-dm domain die? Is the guest running? Is your network > configuration OK to access the VNC server? (it may be easier to just > start by disabling VNC and use a mere PVFB).The domains was started and the vnc port was opened, but after few seconds, the domains died. Follows the final of log file. ******************* KBDFRONT for /local/domain/30/device/vkbd/0 ********** Time offset set 0 evtchn_open() -> 8 xc_evtchn_bind_interdomain(29, 4) = 0 ******************* PCIFRONT for device/pci/0 ********** Failed to read device/pci/0/backend-id. no backend pcilib: minios_init open failed Register xen platform. Done register platform. xs_read_watch() -> /local/domain/0/device-model/29/command dm-command xs_read(/local/domain/0/device-model/29/command): ENOENT xs_read_watch() -> /local/domain/0/device-model/29/logdirty/next-active logdirty xs_read(/local/domain/0/device-model/29/logdirty/key): ENOENT xs_read_watch() -> /local/domain/0/backend/vbd/30/5632/params hdc backend at /local/domain/0/backend/vkbd/30/0 ******************* FBFRONT for /local/domain/30/device/vfb/0 ********** backend at /local/domain/0/backend/vfb/30/0 /local/domain/0/backend/vfb/30/0 connected /local/domain/0/backend/vkbd/30/0 connected ************************** FBFRONT ************************** KBDFRONT Thread "kbdfront" exited. kbd_open(/local/domain/30/device/vkbd/0) -> 9 fb_open(/local/domain/30/device/vfb/0) -> 10 -- Marco Sinhoreli _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Fri 29 Aug 2008 13:54:28 -0300, a écrit :> > So, what''s the problem? :) > > Does the hvmtest-dm domain die? Is the guest running? Is your network > > configuration OK to access the VNC server? (it may be easier to just > > start by disabling VNC and use a mere PVFB). > > The domains was started and the vnc port was opened, but after few > seconds, the domains died. Follows the final of log file.Mmm, nothing strange here, could you check other log files, like in xend.log to know which part actually stopped working at some point? Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hi Samuel, Attached /var/log/xen/xend.log and /var/log/xen/qemu-dm-hvmtest.log. Regards, On Fri, Aug 29, 2008 at 2:02 PM, Samuel Thibault <samuel.thibault@ens-lyon.org> wrote:> Marco Sinhoreli, le Fri 29 Aug 2008 13:54:28 -0300, a écrit : >> > So, what's the problem? :) >> > Does the hvmtest-dm domain die? Is the guest running? Is your network >> > configuration OK to access the VNC server? (it may be easier to just >> > start by disabling VNC and use a mere PVFB). >> >> The domains was started and the vnc port was opened, but after few >> seconds, the domains died. Follows the final of log file. > > Mmm, nothing strange here, could you check other log files, like in > xend.log to know which part actually stopped working at some point? > > Samuel >-- Marco Sinhoreli _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Fri 29 Aug 2008 14:29:52 -0300, a écrit :> [2008-08-29 14:20:29 3620] INFO (XendDomainInfo:1634) Domain has shutdown: name=hvmtest id=33 reason=poweroff.Ok, so that''s not the -dm domain which initiated the shutdown. It''d be probably useful to increase the hypervisor debug level to get the hvmloader messages in xm dmesg. Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hello Samuel,
I increase the debug level of xend. There are two thinks that I looked
in xend.log:
1 - the stub domain was initialized, therefore have no problem with this domain.
2 - After unpause the domain-dm, it has shutdown but I don't understand why.
[2008-09-01 15:06:17 15814] INFO (XendDomain:1172) Domain
hvmtest-dm (25) unpaused.
[2008-09-01 15:06:19 15814] INFO (XendDomainInfo:1634) Domain has
shutdown: name=hvmtest id=24 reason=poweroff.
Attached the xend.log.
Regards,
On Sat, Aug 30, 2008 at 5:37 PM, Samuel Thibault
<samuel.thibault@ens-lyon.org> wrote:> Marco Sinhoreli, le Fri 29 Aug 2008 14:29:52 -0300, a écrit :
>> [2008-08-29 14:20:29 3620] INFO (XendDomainInfo:1634) Domain has
shutdown: name=hvmtest id=33 reason=poweroff.
>
> Ok, so that's not the -dm domain which initiated the shutdown.
It'd
> be probably useful to increase the hypervisor debug level to get the
> hvmloader messages in xm dmesg.
>
> Samuel
>
--
Marco Sinhoreli
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel
Sorry, attached a xend.log with TRACE loglevel. On Mon, Sep 1, 2008 at 3:24 PM, Marco Sinhoreli <msinhore@gmail.com> wrote:> Hello Samuel, > > I increase the debug level of xend. There are two thinks that I looked > in xend.log: > 1 - the stub domain was initialized, therefore have no problem with this domain. > 2 - After unpause the domain-dm, it has shutdown but I don't understand why. > [2008-09-01 15:06:17 15814] INFO (XendDomain:1172) Domain > hvmtest-dm (25) unpaused. > [2008-09-01 15:06:19 15814] INFO (XendDomainInfo:1634) Domain has > shutdown: name=hvmtest id=24 reason=poweroff. > > Attached the xend.log. > > Regards, > > On Sat, Aug 30, 2008 at 5:37 PM, Samuel Thibault > <samuel.thibault@ens-lyon.org> wrote: >> Marco Sinhoreli, le Fri 29 Aug 2008 14:29:52 -0300, a écrit : >>> [2008-08-29 14:20:29 3620] INFO (XendDomainInfo:1634) Domain has shutdown: name=hvmtest id=33 reason=poweroff. >> >> Ok, so that's not the -dm domain which initiated the shutdown. It'd >> be probably useful to increase the hypervisor debug level to get the >> hvmloader messages in xm dmesg. >> >> Samuel >> > > > > -- > Marco Sinhoreli >-- Marco Sinhoreli _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Mon 01 Sep 2008 15:24:10 -0300, a écrit :> I increase the debug level of xend.I was referring to the debug level of the hypervisor. Add guest_loglvl=all to your hypervisor command line, and look for HVM: lines in xm dmesg. Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hi Samuel, I found the problem: In qemu start, the hard disk is the default bootable device. (XEN) HVM30: HVM Loader (XEN) HVM30: Detected Xen v3.3.0 (XEN) HVM30: CPU speed is 2810 MHz (XEN) HVM30: Multiprocessor initialisation: (XEN) HVM30: - CPU0 ... 36-bit phys ... fixed MTRRs ... var MTRRs ... done. (XEN) HVM30: Writing SMBIOS tables ... (XEN) HVM30: Loading ROMBIOS ... (XEN) HVM30: 9452 bytes of ROMBIOS high-memory extensions: (XEN) HVM30: Relocating to 0x1fff8800-0x1fffacec ... done (XEN) irq.c:235: Dom30 PCI link 0 changed 0 -> 5 (XEN) HVM30: PCI-ISA link 0 routed to IRQ5 (XEN) irq.c:235: Dom30 PCI link 1 changed 0 -> 10 (XEN) HVM30: PCI-ISA link 1 routed to IRQ10 (XEN) irq.c:235: Dom30 PCI link 2 changed 0 -> 11 (XEN) HVM30: PCI-ISA link 2 routed to IRQ11 (XEN) irq.c:235: Dom30 PCI link 3 changed 0 -> 5 (XEN) HVM30: PCI-ISA link 3 routed to IRQ5 (XEN) HVM30: pci dev 01:2 INTA->IRQ10 (XEN) HVM30: pci dev 03:0 INTA->IRQ5 (XEN) HVM30: pci dev 04:0 INTA->IRQ5 (XEN) HVM30: pci dev 02:0 bar 10 size 02000000: f0000008 (XEN) HVM30: pci dev 03:0 bar 14 size 01000000: f2000008 (XEN) HVM30: pci dev 02:0 bar 14 size 00001000: f3000000 (XEN) HVM30: pci dev 03:0 bar 10 size 00000100: 0000c001 (XEN) HVM30: pci dev 04:0 bar 10 size 00000100: 0000c101 (XEN) HVM30: pci dev 04:0 bar 14 size 00000100: f3001000 (XEN) HVM30: pci dev 01:1 bar 20 size 00000010: 0000c201 (XEN) HVM30: Creating MP tables ... (XEN) HVM30: Loading Cirrus VGABIOS ... (XEN) HVM30: Loading PXE ROM ... (XEN) HVM30: - Manufacturer: Etherboot (XEN) HVM30: - Product name: rtl8139.zrom 5.4.2 (GPL) etherboot.org (XEN) HVM30: Loading ACPI ... (XEN) HVM30: - Lo data: 000ea000-000ea03f (XEN) HVM30: - Hi data: 1fff7400-1fff87df (XEN) HVM30: BIOS map: (XEN) HVM30: c0000-c8bff: VGA BIOS (XEN) HVM30: d0000-d7fff: Etherboot ROM (XEN) HVM30: e9000-e9140: SMBIOS tables (XEN) HVM30: f0000-fffff: Main BIOS (XEN) HVM30: Invoking ROMBIOS ... (XEN) HVM30: rombios.c,v 1.138 2005/05/07 15:55:26 vruppert Exp $ (XEN) stdvga.c:147:d30 entering stdvga and caching modes (XEN) HVM30: HVMAssist BIOS, 1 cpu, $Revision: 1.138 $ $Date: 2005/05/07 15:55:26 $ (XEN) HVM30: (XEN) HVM30: ata0-0: PCHS=16383/16/63 translation=lba LCHS=1024/255/63 (XEN) HVM30: ata0 master: QEMU HARDDISK ATA-7 Hard-Disk (10240 MBytes) (XEN) HVM30: ata0 slave: Unknown device (XEN) HVM30: ata1 master: QEMU CD-ROM ATAPI-4 CD-Rom/DVD-Rom (XEN) HVM30: ata1 slave: Unknown device (XEN) HVM30: (XEN) HVM30: (XEN) HVM30: (XEN) HVM30: Press F10 to select boot device. (XEN) HVM30: Booting from Hard Disk... (XEN) HVM30: Boot from Hard Disk failed: not a bootable disk (XEN) HVM30: FATAL: No bootable device. (XEN) hvm.c:727:d30 All CPUs offline -- powering off. I set in config file to boot from cdrom setting the variable boot=''dc'' but it isn''t working. Well, after start the virtual machine, I open fastlly the vncviewer to connect to virtual machine and I pressed the F10 key to select the #3 option to boot from CDrom and It worked. (XEN) HVM32: HVM Loader (XEN) HVM32: Detected Xen v3.3.0 (XEN) HVM32: CPU speed is 2810 MHz (XEN) HVM32: Multiprocessor initialisation: (XEN) HVM32: - CPU0 ... 36-bit phys ... fixed MTRRs ... var MTRRs ... done. (XEN) HVM32: Writing SMBIOS tables ... (XEN) HVM32: Loading ROMBIOS ... (XEN) HVM32: 9452 bytes of ROMBIOS high-memory extensions: (XEN) HVM32: Relocating to 0x1fff8800-0x1fffacec ... done (XEN) irq.c:235: Dom32 PCI link 0 changed 0 -> 5 (XEN) HVM32: PCI-ISA link 0 routed to IRQ5 (XEN) irq.c:235: Dom32 PCI link 1 changed 0 -> 10 (XEN) HVM32: PCI-ISA link 1 routed to IRQ10 (XEN) irq.c:235: Dom32 PCI link 2 changed 0 -> 11 (XEN) HVM32: PCI-ISA link 2 routed to IRQ11 (XEN) irq.c:235: Dom32 PCI link 3 changed 0 -> 5 (XEN) HVM32: PCI-ISA link 3 routed to IRQ5 (XEN) HVM32: pci dev 01:2 INTA->IRQ10 (XEN) HVM32: pci dev 03:0 INTA->IRQ5 (XEN) HVM32: pci dev 04:0 INTA->IRQ5 (XEN) HVM32: pci dev 02:0 bar 10 size 02000000: f0000008 (XEN) HVM32: pci dev 03:0 bar 14 size 01000000: f2000008 (XEN) HVM32: pci dev 02:0 bar 14 size 00001000: f3000000 (XEN) HVM32: pci dev 03:0 bar 10 size 00000100: 0000c001 (XEN) HVM32: pci dev 04:0 bar 10 size 00000100: 0000c101 (XEN) HVM32: pci dev 04:0 bar 14 size 00000100: f3001000 (XEN) HVM32: pci dev 01:1 bar 20 size 00000010: 0000c201 (XEN) HVM32: Creating MP tables ... (XEN) HVM32: Loading Cirrus VGABIOS ... (XEN) HVM32: Loading PXE ROM ... (XEN) HVM32: - Manufacturer: Etherboot (XEN) HVM32: - Product name: rtl8139.zrom 5.4.2 (GPL) etherboot.org (XEN) HVM32: Loading ACPI ... (XEN) HVM32: - Lo data: 000ea000-000ea03f (XEN) HVM32: - Hi data: 1fff7400-1fff87df (XEN) HVM32: BIOS map: (XEN) HVM32: c0000-c8bff: VGA BIOS (XEN) HVM32: d0000-d7fff: Etherboot ROM (XEN) HVM32: e9000-e9140: SMBIOS tables (XEN) HVM32: f0000-fffff: Main BIOS (XEN) HVM32: Invoking ROMBIOS ... (XEN) HVM32: rombios.c,v 1.138 2005/05/07 15:55:26 vruppert Exp $ (XEN) stdvga.c:147:d32 entering stdvga and caching modes (XEN) HVM32: HVMAssist BIOS, 1 cpu, $Revision: 1.138 $ $Date: 2005/05/07 15:55:26 $ (XEN) HVM32: (XEN) HVM32: ata0-0: PCHS=16383/16/63 translation=lba LCHS=1024/255/63 (XEN) HVM32: ata0 master: QEMU HARDDISK ATA-7 Hard-Disk (10240 MBytes) (XEN) HVM32: ata0 slave: Unknown device (XEN) HVM32: ata1 master: QEMU CD-ROM ATAPI-4 CD-Rom/DVD-Rom (XEN) HVM32: ata1 slave: Unknown device (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: Press F10 to select boot device. (XEN) HVM32: Key pressed: 0044 (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: Select boot device (XEN) HVM32: (XEN) HVM32: 1. Floppy (XEN) HVM32: 2. Hard drive (XEN) HVM32: 3. CD-ROM (XEN) HVM32: 4. Network (XEN) HVM32: (XEN) HVM32: (XEN) HVM32: Currently selected: 3 (XEN) HVM32: Key pressed: 001C (XEN) HVM32: Booting from CD-Rom... (XEN) stdvga.c:151:d32 leaving stdvga (XEN) common.c:3541:d33 tracking VRAM f0000 - f00f0 There is another way to set the default bootable device in conffile? Something was changed? What changed? Best regards, -- Marco Sinhoreli _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Marco Sinhoreli, le Mon 01 Sep 2008 18:09:09 -0300, a écrit :> I set in config file to boot from cdrom setting the variable boot=''dc''In hvmtest, right? If so, can you check that the qemu-dm-hvmtest.log shows "-boot" "d"? Samuel _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel