В Пт, 29/05/2015 в 17:20 +0300, Andrew Vagin пишет:
Acked-by: Andrew Vagin ava...@odin.com
I'm agree that we need to remove this function, but I don't know how it
fixes the bug.
tcp_v4_kill_ve_sockets() is called when all processes are killed,
and their sockets have been already replaced
В Пн, 01/06/2015 в 15:16 +0300, Vasily Averin пишет:
On 01.06.2015 14:31, Kirill Tkhai wrote:
В Пт, 29/05/2015 в 17:20 +0300, Andrew Vagin пишет:
Acked-by: Andrew Vagin ava...@odin.com
I'm agree that we need to remove this function, but I don't know how it
fixes the bug.
On 01.06.2015 14:31, Kirill Tkhai wrote:
В Пт, 29/05/2015 в 17:20 +0300, Andrew Vagin пишет:
Acked-by: Andrew Vagin ava...@odin.com
I'm agree that we need to remove this function, but I don't know how it
fixes the bug.
tcp_v4_kill_ve_sockets() is called when all processes are killed,
and
В Пн, 01/06/2015 в 17:25 +0300, Vasily Averin пишет:
On 01.06.2015 15:46, Kirill Tkhai wrote:
В Пн, 01/06/2015 в 15:16 +0300, Vasily Averin пишет:
On 01.06.2015 14:31, Kirill Tkhai wrote:
В Пт, 29/05/2015 в 17:20 +0300, Andrew Vagin пишет:
Acked-by: Andrew Vagin ava...@odin.com
I'm
Any plans to sync these limits/counters with user beancounters?
___
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel
On 06/01/2015 07:01 PM, Vladimir Davydov wrote:
On Fri, May 29, 2015 at 05:19:15PM +0300, Pavel Emelyanov wrote:
Packet sockets have incoming queue of packets that is only limited
with per-socket wmem buffer. Strictly speaking we should sum up
all the queues and charge them into kmem once new
On Mon, Jun 01, 2015 at 07:11:51PM +0300, Pavel Emelyanov wrote:
On 06/01/2015 07:02 PM, Vladimir Davydov wrote:
Any plans to sync these limits/counters with user beancounters?
For TCP it would be too hard. Current tcp mm doesn't distinguish in mem
from out mem, so we have no easy way to
On Mon, Jun 01, 2015 at 07:14:24PM +0300, Pavel Emelyanov wrote:
On 06/01/2015 07:01 PM, Vladimir Davydov wrote:
On Fri, May 29, 2015 at 05:19:15PM +0300, Pavel Emelyanov wrote:
Packet sockets have incoming queue of packets that is only limited
with per-socket wmem buffer. Strictly speaking
Author: Vladimir Davydov
Email: vdavy...@parallels.com
Subject: sched: make nr_cpus limit support hierarchies
Date: Wed, 26 Nov 2014 17:29:31 +0300
Before the patch nr_cpus limiter was completely unaware of hierarchies,
i.e. creation of a cpu sub-cgroup in a container would result in its
tasks
On Fri, May 29, 2015 at 05:19:15PM +0300, Pavel Emelyanov wrote:
Packet sockets have incoming queue of packets that is only limited
with per-socket wmem buffer. Strictly speaking we should sum up
all the queues and charge them into kmem once new packet arrives,
but this will result in huge
On 06/01/2015 07:02 PM, Vladimir Davydov wrote:
Any plans to sync these limits/counters with user beancounters?
For TCP it would be too hard. Current tcp mm doesn't distinguish in mem
from out mem, so we have no easy way to show them separately. For the
rest... Well, have you changed your mind
Thank you, this approach looks way better and in-line with current vzctl
ideas!
Can you please:
1. Use git format-patch
2. Provide some description (in the patch commit message) about why you
need it etc.
An example of such is
https://src.openvz.org/projects/OVZL/repos/vzctl/commits/4772f7e
On 01.06.2015 15:46, Kirill Tkhai wrote:
В Пн, 01/06/2015 в 15:16 +0300, Vasily Averin пишет:
On 01.06.2015 14:31, Kirill Tkhai wrote:
В Пт, 29/05/2015 в 17:20 +0300, Andrew Vagin пишет:
Acked-by: Andrew Vagin ava...@odin.com
I'm agree that we need to remove this function, but I don't know
On 06/01/2015 07:23 PM, Vladimir Davydov wrote:
On Mon, Jun 01, 2015 at 07:11:51PM +0300, Pavel Emelyanov wrote:
On 06/01/2015 07:02 PM, Vladimir Davydov wrote:
Any plans to sync these limits/counters with user beancounters?
For TCP it would be too hard. Current tcp mm doesn't distinguish in
14 matches
Mail list logo