0

I am having xen 3.4.2 running with 15 vm (pv+hvm) on a host with 256 Gb RAM and 96 core CPU.

But recently my host crashed on debug log its showing

translating ffff83183fcb0000 with CR3 100ae42000 and 4 levels of page table. 

after so many similar lines its showing

cannot translate address 0 < ffff830000000000 without cr3 

From my understanding of xen pv is that,

hypervisor gives pv direct access to physical RAM

but hypervisor cross checks all the call to physical memory rather than using shadow pages.

So it has less overhead in virtual memory to physical conversion,as its aware about actual mapping.

But in case of HVM hypervisor needs to translate guests memory to physical ram.

So can anyone explain me from above translation that its for hvm ram translation hypervisor is doing or will it happen in pv also?

and on crash.log its showing

(XEN) grant_table.c:1408:d0 dest domain 452 dying (XEN) p2m_pod_cache_get: Breaking up superpage. (XEN) mm.c:741:d421 Non-privileged (421) attempt to map I/O space 00000000 (XEN) mm.c:741:d421 Non-privileged (421) attempt to map I/O space 000000f0 (XEN) mm.c:741:d352 Non-privileged (352) attempt to map I/O space 00000000 (XEN) mm.c:741:d352 Non-privileged (352) attempt to map I/O space 000000f0 (XEN) mm.c:741:d249 Non-privileged (249) attempt to map I/O space 00000000 (XEN) mm.c:741:d249 Non-privileged (249) attempt to map I/O space 000000f0 (XEN) grant_table.c:1408:d0 dest domain 450 dying 

This is the second crash in a month.

I have seen many system programming related questions here thats why i posted it here.

1 Answer 1

0

Xen doesn't give direct access to RAM, instead it allows the Physical Volume to use physical RAM for page table updating but under the supervision of Xen.

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.