[
https://issues.apache.org/jira/browse/HTTPCLIENT-1310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13576237#comment-13576237
]
Martin Meinhold commented on HTTPCLIENT-1310:
---------------------------------------------
Thanks for your feedback, it is very appreciated.
"Did you test if this patch improves the performance of the cache in the case
when the remote server is down?"
The change can reduce the number of requests send to the remote server
significantly. Assuming the remote server is down and the background validation
is required, ATM every request will trigger the background validation. In our
scenario we want performance over consistency, thus we are more interested in
limiting the number of requests (so lower the resource consumption) with the
drawback on not being up-to-date when the server is back. And yes, I've tested
successfully the patch.
"From the client point of view, as re-validation is done in the background, I
guess it should not make any difference."
Correct.
"Why do you want to do this?"
We use the HttpClient as a transparent cache which stores documents from remote
servers. Thus we can access these documents very frequently without creating
requests to the remote server. Also the background validation is very handy in
that situation as it avoids blocking a thread while waiting on the remote
server response.
As soon as the remote server goes down or returns a 5xx status code, every
request will trigger a network request to the remote server (which is of course
the job of the HttpClient). As described above, it is acceptable for us to work
with a stale document and try the background validation from time to time (e.g.
once a minute or every 30 seconds). But the current implementation doesn't
allow to change the frequency when the background validation is run.
I understand that as a client, I can check for the existance of the Warning
header. But IMHO the value of the Warning ('110 localhost "Response is stale"')
is not very machine readable and looks too fragile to me to build my error
handling on top of it. And even if I know that I want to avoid the next
background validation - how can I do that? Is there an easy way to retrieve the
cached document from the HttpClient without triggering a background validation?
The patch would make the information that the last x background validation
requests have failed more explicit. By extending the
ExponentialBackOffSchedulingStrategy consuming clients can even implement their
own algorithm. While testing this again today I discovered that due to
AsynchronousRevalidationRequest being package protected it may be handy to
extract some code in an abstract class ...
Again, please feel free to post any feedback, concerns and comments.
> Allow background validation to optionally back off after a number of failed
> requests
> ------------------------------------------------------------------------------------
>
> Key: HTTPCLIENT-1310
> URL: https://issues.apache.org/jira/browse/HTTPCLIENT-1310
> Project: HttpComponents HttpClient
> Issue Type: Improvement
> Components: Cache
> Reporter: Martin Meinhold
> Attachments:
> 0001-HTTPCLIENT-1310-Allow-clients-to-change-the-used-sch.patch
>
>
> We are successfully using the background validation to asynchronously update
> cache entries while returning a stale document (stale-while-revalidate cache
> control header). Also in case an error has happened, the stale document is
> used (stale-if-error cache control header). Works perfectly. Guys - great
> work you made this happen.
> Now the tricky part: as soon there is an issue like e.g. the remote server is
> down, the stale-if-error header prevent the cache from being updated (which
> of course is the intention of that header). But this also means, that code
> using the HttpClient has no way to discover that there was an issue. So every
> following request will get that stale document but also trigger a background
> revalidation.
> As an improvement it should be possible that the background validation backs
> off after a certain amount of failed requests. This should be optional and
> not the default.
> I want to contribute some code we already have working on a 4.2 branch. The
> central idea is to vary the scheduling strategy the AsynchronousValidation
> uses to estimate _when_ the background validation of a certain request should
> happen. Of course, the default would be immediately.
> In fact this would move code currently submitting tasks to the executor from
> the AsynchronousValidation into a separate class. Thus the
> AsynchronousValidation would become kind of a director role by simply
> enqueuing next requests and keeping track which of them failed and which were
> successful. A strategy could - based on the failure count - execute them
> immediately or later. Again, to clarify: the default behaviour would be to
> execute every incoming background validation request immediately regardless
> of the error count.
> What do you think?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
