Re: [Xen-users] After Update: NUMA placement failed, performance might be affected

[Dario Faggioli <dario.faggioli@xxxxxxxxxx>]

On Thu, 2017-01-05 at 11:19 +0100, Helmut Schlang wrote:
> Hello.
> 
> After I updated my Xen server that's running on debian jessie, I can
> not
> create PV nor HVM guests anymore. 
>
Updated from what to what? (The latter is probably 4.4).

> Everything was working perfectly fine
> before the update. What does "NUMA placement failed" even mean?
> 
It's a warning that libxl tried to figure out what was the best
possible distribution of guests' memory on the host's NUMA node, but
that failed for some reason.

As said, it's only a warning, and optimal placement failure won't
prevent guest creation, if there is enough memory for that, in total,
on the host.

> root@hpx01:~# xl -vvvv create /xen/windows.cfg
> Parsing config from /xen/windows.cfg
> libxl: debug: libxl_create.c:1347:do_domain_create: ao 0x8860cd0:
> create: how=(nil) callback=(nil) poller=0x8860ee0
> libxl: debug: libxl_device.c:296:libxl__device_disk_set_backend: Disk
> vdev=hda spec.backend=unknown
> libxl: debug: libxl_device.c:331:libxl__device_disk_set_backend: Disk
> vdev=hda, using backend phy
> libxl: debug: libxl_device.c:296:libxl__device_disk_set_backend: Disk
> vdev=hdb spec.backend=unknown
> libxl: debug: libxl_device.c:331:libxl__device_disk_set_backend: Disk
> vdev=hdb, using backend phy
> libxl: debug: libxl_create.c:799:initiate_domain_create: running
> bootloader
> libxl: debug: libxl_bootloader.c:321:libxl__bootloader_run: not a PV
> domain, skipping bootloader
> libxl: debug: libxl_event.c:618:libxl__ev_xswatch_deregister: watch
> w=0x88610d0: deregister unregistered
> libxl: notice: libxl_numa.c:494:libxl__get_numa_candidate: NUMA
> placement failed, performance might be affected
>
So, this is the warning. And it's not here that things are failing.
That happens...

> xc: detail: elf_parse_binary: phdr: paddr=0x100000 memsz=0x59658
> xc: detail: elf_parse_binary: memory: 0x100000 -> 0x159658
> xc: detail: VIRTUAL MEMORY ARRANGEMENT:
>   Loader:        0000000000100000->0000000000159658
>   Modules:       0000000000000000->0000000000000000
>   TOTAL:         0000000000000000->00000000fd500000
>   ENTRY ADDRESS: 0000000000100648
> xc: detail: Failed allocation for dom 6: 160 extents of order 0
> xc: error: Could not allocate memory for HVM guest. (16 = Device or
> resource busy): Internal error
> libxl: error: libxl_dom.c:655:libxl__build_hvm: hvm building failed
> libxl: error: libxl_create.c:1024:domcreate_rebuild_done: cannot
> (re-)build domain: -3
>
...here! And "Failed allocating for dom yyy" and "Could not allocate
memory for xxx" are the actual error messages.

How much memory does the guest have, how much RAM do you have on the
host and how much of it is free before trying to create this new guest
(you can, for instance, post the output of `xl info -n').

Regards,
Dario
-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
https://lists.xen.org/xen-users