Hi,
When using the %t option in the LogFormat directive, when is the
timestamp generated? Is it when the server receives the request, or is
it when the CLIENT sends the request.

I am receiving requests in a strange order (not always ascending time
order), and was wondering where to start searching for the problem.

Thanks
Codemwnci

On 1 Mar 2005 01:57:59 -0800,
"codemwnci" <ellisw@aisl.uk.com> posted:

> When using the %t option in the LogFormat directive, when is the
> timestamp generated? Is it when the server receives the request, or is
> it when the CLIENT sends the request.

I would say when the server receives it. How would it know when it was
sent? But read further on.

> I am receiving requests in a strange order (not always ascending time
> order), and was wondering where to start searching for the problem.

I've seen that sort of thing, too. It's odd to see requests for things
referenced in a page logged before the actual page. Which makes me wonder
if the server logs *don't* show the times for when a request was *made*,
but when it *finished* being served. Which would make sense, as how else
would it log the bytes that were sent (remember, a download can be aborted,
and that'll be logged instead, although differently).

e.g. 1. Page requested, and it starts downloading.
2. As it downloads, CSS files are requested. They're short, so
they're logged now.
3. Page is still downloading, and some tiny images are downloaded,
and logged.
4. Page finishes downloading, and is logged.
5. Some other larger images referenced by the page are downloaded
and logged.

--
If you insist on e-mailing me, use the reply-to address (it's real but
temporary). But please reply to the group, like you're supposed to.

This message was sent without a virus, please delete some files yourself.

On 1 Mar 2005 01:57:59 -0800, in comp.infosystems.www.servers.unix,
"codemwnci" <ellisw@aisl.uk.com> wrote:

>Hi,
>When using the %t option in the LogFormat directive, when is the
>timestamp generated? Is it when the server receives the request, or is
>it when the CLIENT sends the request.
>
>I am receiving requests in a strange order (not always ascending time
>order), and was wondering where to start searching for the problem.
>

I've examined the source in both Apache 1 and Apache 2 - I wanted to
see why I could log page serve times in milliseconds in Apache 2 and
only seconds in Apache 1. From what I see, Apache 1 just uses the
time that the log entry is written - the end of the request. Apache 2
has more complex code and several IFDEFs that allow multiple options -
it appears that the default is to use the request time stored in the
request structure - ICBW.

HTH,
Jim