> On 27/04/2016, at 12:48 PM, Jerry Jelinek <[email protected]> wrote:
>
> A few things. Can you provide more information about what did not work when
> you set the zone's memory cap?
It did not cap the memory, the zone continued to allocate. Ah, it turns out I
have this on another window:
[root@i7 ~]# zonememstat
ZONE RSS(MB) CAP(MB) NOVER POUT(MB)
global 98 - - -
7302f2d4-a0a4-4686-9497-b9be94bf16d6 6159 4096 0 0
I'm sure I read somewhere (although it may apply to a different capping method)
that it's actually supposed to let it go over cap if there's plenty of spare
capacity - presumably clawing it back by paging stuff out. While running
'steady state', vmstat in the global zone reported
0 0 0 12062508 3638504 550 98371 0 0 0 0 0 0 0 0 0 3799 50334 1137 83 13 4
0 0 0 12301720 3878096 2438 112891 0 0 0 0 0 0 0 0 0 4495 46501 1562 83 12 5
0 0 0 11981296 3557276 17 97493 0 0 0 0 0 0 0 0 36 4336 19054 1906 85 10 5
0 0 0 11616724 3192284 559 96247 0 0 0 0 0 0 0 0 0 3797 3475 1206 86 10 4
0 0 0 11679048 3255508 2001 104195 0 0 0 0 0 0 0 0 0 4370 52181 1581 83 12 5
0 0 0 11348032 2925224 0 76616 0 0 0 0 0 0 0 0 0 3688 11573 1299 87 7 5
kthr memory page disk faults cpu
r b w swap free re mf pi po fr de sr lf rm s0 s1 in sy cs us sy id
0 0 0 11117224 2694352 1079 66857 0 0 0 0 0 0 0 0 0 3547 3901 1206 88 8 4
So what looks, to me, like some frantic reclaiming of pages. Note as well the
zero runnable threads.
> If you have questions about how the zone memory cap is expected to work, I'll
> try to answer them.
OK, ah, I've set a cap in the zone's config but it's not being adhered to. The
machine eventually runs out of free ram and stops. I've seen it recover once or
twice so I assume it's in the process of recovering now, just not getting
anywhere.
My biggest priority is ensuring a zone can't take down another zone, and being
able to log in to global is a base requirement, really.
> Second, although the design center for smartos does not include the kind of
> customization you've done in the global zone, I would be interested in
> debugging the live lock you're in.
I'll try to reproduce on VMWare and if that's not a goer then I'll reproduce on
hardware without any customisations.
> If you're willing to provide a panic dump, I'd like to take a look at it.
I don't think it's panic'ed. I can still ping it, for instance.
I'll see what I can do. It's obviously not a case of me misunderstanding
something.
Thanks,
Dave
-------------------------------------------
smartos-discuss
Archives: https://www.listbox.com/member/archive/184463/=now
RSS Feed: https://www.listbox.com/member/archive/rss/184463/25769125-55cfbc00
Modify Your Subscription:
https://www.listbox.com/member/?member_id=25769125&id_secret=25769125-7688e9fb
Powered by Listbox: http://www.listbox.com
