httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "S.A." <>
Subject Re: [users@httpd] Serving partial data of in-memory common data set
Date Wed, 29 Jul 2009 18:18:17 GMT
> ...
> > We started with 30 concurrent users and there was no
> trouble,
> > but when the next batch of 70 users hit concurrently,
> we could
> > not serve all the users.
> > 
> Just a note for now.  The above is an important part,
> and maybe you should have started your initial post with
> that information.
> It shows that you have a running system, that you have
> tried your application, and that you have already seen a
> problem.
> I know it is a matter of presentation, but it puts
> everything in a different light and is more likely to
> attract attention and helpful responses, than the way you
> presented your initial post, which looked more like you were
> looking for theoretical answers to hypothetical questions.

May be I should have. I was debating between a long post
and a summary and in the interest of time, tried to post
what I was looking for. 

> > 
> > We have about about 10 set of 50 pages and each page
> containing
> > varying number of images (on an average 45 images of
> about 2KB
> > and about 50 or so 0.5KB images) and an occasional
> multi-media
> > file. For this discussion, we can ignore the
> multi-media file. 
> > The page content and the sequence of page presentation
> is
> > user dependent, but for a group of users this 50 page
> set
> > is constant. With each request we do updates on 3
> mysql tables.
> > 
> The last sentence is also something worth investigating.
> Such as : is the bottleneck at the level of the images, or
> at the level of the mysql accesses ?

The reason why I am not suspecting mysql was that the mysql
log does indicate that it is getting all the requests and
it is servicing them. As I have stated before, some of the
users though are not getting images.

> ...
> > Our eventual objective is to serve about 200,000 users
> (of
> > course not with our existing hardware) and we are
> looking
> > at various options.
> That is also a good number to give, as it gives a measure
> of the goal.
> ...
> > 
> > We have definitely a problem and looking at various
> options
> > to resolve it.
> My first reaction to the above would be that yes, you do
> have a problem.  If you cannot serve the kind of
> content you mention for 70 users, with the hardware you
> mention, then something seems definitely amiss.
> I cannot imagine that you could have a problem serving 70
> clients simultaneously, with pages containing 45 images of 2
> KB each.
> My gut feeling tells me that you may have a concurrency
> problem at the level of your mysql accesses.
> Would it be possible, for instance, to replace the mysql
> accesses by some static content fetched from disk (for
> example, 1 distinct text file for each of your client-ids,
> the name being client-id driven) ?
> And then see if you have a problem ?

There is very little data that gets served to the user from
mysql. We use mysql to maintain statistics on what content
user is accessing and the sequence in which the info is
presented as the order matters to us. So, it is predominantly
updates that we do to the mysql. Of course, updates are more
latency intense than reads, yet those operations seem to
get thru.

> The reason for that gut feeling is that you wrote somewhere
> that you can serve up to a certain number of clients fine,
> but when increasing the number, you hit a wall.

We have looked at mysql aspect and will continue to look
at it. Hardware is another issue we need to resolve and
then may be memcached.

Thank you again for your responses.


The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:> for more info.
To unsubscribe, e-mail:
   "   from the digest:
For additional commands, e-mail:

View raw message