On Tue, Nov 4, 2014 at 7:36 PM, Rajalakshmi Iyer wrote:
> Ok. Thanks.
>
> Does this situation sound plausible - Keep alive connections are not being
> used effectively by client applications (i.e. subsequent requests on
> keep-alive connections are not frequent enough to keep Apache workers busy
>
On Tue, Nov 04, 2014 at 03:58:25PM +, Rajalakshmi Iyer wrote:
> The server has a KeepAliveTimeout of 120 seconds.
That is probably the longest KeepAliveTimeout I have seen. Is there some
particular reason that you have it so high? In 2.4 the default value is
5 seconds, which ought to be a rea
Am 04.11.2014 um 20:36 schrieb Rajalakshmi Iyer:
Ok. Thanks.
Does this situation sound plausible - Keep alive connections are not
being used effectively by client applications (i.e. subsequent requests
on keep-alive connections are not frequent enough to keep Apache workers
busy all the time, re
On Tue, Nov 4, 2014 at 2:36 PM, Rajalakshmi Iyer wrote:
> Does this situation sound plausible - Keep alive connections are not being
> used effectively by client applications (i.e. subsequent requests on
> keep-alive connections are not frequent enough to keep Apache workers busy
> all the time, r
Ok. Thanks.
Does this situation sound plausible - Keep alive connections are not being
used effectively by client applications (i.e. subsequent requests on
keep-alive connections are not frequent enough to keep Apache workers busy
all the time, resulting in low CPU utilisation). At the same time,
On Tue, Nov 4, 2014 at 2:17 PM, Rajalakshmi Iyer wrote:
> Could you also elaborate on the statement --- With event, keep alive
> connections don't tie up a thread between requests. Does that mean a thread
> is allotted per connection, rather than per request in case of keep-alive
> connections (th
So we have a situation where keep=alive connections are not being used
effectively by client applications (i.e. subsequent requests on keep-alive
connections are not frequent enough to keep Apache workers busy all the
time, resulting in low CPU utilisation). At the same time, because there
are alre
On Tue, Nov 4, 2014 at 11:23 AM, Eric Covener wrote:
> No, with event, keepalive connections don't tie up a thread between requests.
Sorry, I see what you mean. Yes, each process will stop calling accept
out of fear that too many keepalive connections may become active (and
hence be queued). Each
No, with event, keepalive connections don't tie up a thread between requests.
On Tue, Nov 4, 2014 at 10:58 AM, Rajalakshmi Iyer wrote:
> Lets say the server is running with a max of 720 connections
> (AsyncRequestWorkerFactor + 1) * MaxRequestWorkers (in this case 3 * 240 =
> 720)
>
> No new conn
Lets say the server is running with a max of 720 connections *(*
AsyncRequestWorkerFactor* + 1) * **MaxRequestWorkers *(in this case 3 * 240
= 720)
No new connections will be accepted by the server at this point, until the
current connections is < *ThreadsPerChild +
(AsyncRequestWorkerFactor * num
On Tue, Nov 4, 2014 at 9:40 AM, Rajalakshmi Iyer wrote:
> In case of an event MPM, unlike the worker MPM, can it be expected that an
> idle worker thread starts working on a new connection or new requests on
> another existing connection?
>
> Or is the worker thread tied to a connection and will
ct: Re: [users@httpd] Help needed with event MPM configuration
In case of an event MPM, unlike the worker MPM, can it be expected that an
idle worker thread starts working on a new connection or new requests on
another existing connection?
Or is the worker thread tied to a connection and wil
In case of an event MPM, unlike the worker MPM, can it be expected that an
idle worker thread starts working on a new connection or new requests on
another existing connection?
Or is the worker thread tied to a connection and will only work on requests
on that connection till the connection is cl
On Tue, Nov 4, 2014 at 9:10 AM, Rajalakshmi Iyer wrote:
> We might be able to increase the number of threads available to handle new
> connections, but it seems like existing threads are doing nothing (CPU
> utilization is quite low). How can that be explained?
Threads in Apache work on a single
The CPU consumed is the load required to fulfil 240 requests simultaneously
while keeping the rest queued.
As Eric said, If you want to use more of your available CPU you need to
increase MaxRequestWorkers to 720 or more.
Also ServerLimit and ThreadsPerChild as their multiple is currently 240 so
users@httpd.apache.org
Subject: Re: [users@httpd] Help needed with event MPM configuration
Thanks Eric.
We might be able to increase the number of threads available to handle new
connections, but it seems like existing threads are doing nothing (CPU
utilization is quite low). How can that be exp
Thanks Eric.
We might be able to increase the number of threads available to handle new
connections, but it seems like existing threads are doing nothing (CPU
utilization is quite low). How can that be explained?
Regards
Raj
On Tue, Nov 4, 2014 at 2:05 PM, Eric Covener wrote:
> > Assuming the
> Assuming the server is always under heavy load, the 240 worker threads are
> busy working on 240 requests from 240 established connections. As soon as
> they finish their work, they will be working on the subsequent requests on
> the same connection.
>
> There are 720 connections that have bee
18 matches
Mail list logo
