I like it.
Do you have any more info on the Concurrency
issues? I have a pretty good debugging setup going here and I'm
about to upgrade to rails 2.2.2.
I can try to help track em down
if you can provide me with more intel. (or a
link)
Jay
Apr 22, 2009 01:22:48 AM,
[email protected] wrote:
AD wrote:
>
Just wanted to let everyone know I created a new section on the
WIKI
> for rails mult-threading best practices. We have been
getting hit
> with a few issues and as this is one of the key
benefits of
> Jruby/Rails 2.2+ I thought it would be good to
consolidate
> issues/recommendations/etc on the Wiki. Would
definitely appreciate
> any feedback/recommendations that people
have on this subject.
>
> I know one of the key things we
are toying with now is how many
> "MaxThreads" to set in Tomcat as
well as looking for major gotchas in
> the controller code that
can cause these Concurrency errors we have
> been seeing.
>
> http://kenai.com/projects/jruby/pages/RailsMulti-ThreadingBestPractices/
That's
great! I'll review it and try to add anything I think is missing
later today. Thank you for creating this!
-
Charlie
---------------------------------------------------------------------
To
unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
Thanks Jay. The trouble is it seems to happen really only under our
load tests, so not sure what is happening there. When it does happen
we get the “InvalidCacheDescendants” error but no real backtrace in
the code to help determine the root of the problem.
Would you mind sharing some of your debugging techniques? We are
about to modify the exception_notifier plugin gem to provide some more
debugging output.
I wish i can give you a URL directly but the site is pretty secured
right now unfortunately. I will certainly post information as I
discover it.
Adam
On Wed, Apr 22, 2009 at 11:04 AM, Jay McGaffigan
[email protected] wrote:
I know one of the key things we are toying with now is how many
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
--------------------------------------------------------------------- To
unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
no prob. I am in the process of upgrading to rails 2.3.2 and jruby
1.2. We also have our own heavy load tests. So maybe I can
reproduce the issue you are seeing. Are you using SSL?
Typically we use JConsole and YourKit And we run against glassfish and
tomcat to try and “see” differences. Yourkit instruments into GF and
Tomcat and allows us to view thread information. Sometimes it’s hard
to read the info but sometimes we stumble upon resource leaks etc.
That’s how we found the SSL session problems we were seeing.
Feel free to reach out even if you want to just talk shop. Sometimes
talking about what you are doing with other sets of eyes/ears helps
spring new ideas.
Jay
On Wed, Apr 22, 2009 at 1:52 PM, AD [email protected] wrote:
right now unfortunately. I will certainly post information as I
for rails mult-threading best practices. We have been getting hit
http://kenai.com/projects/jruby/pages/RailsMulti-ThreadingBestPractices/
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
Thanks Jay, our next step is to get deeper into JConsole during a load
test to see what is happening. I am just not sure if we will be able
to get an accurate point in time snap of the thread that caused the
error, not to mention we have hundres of active threads, so not sure
if you have any recommendations on how to help track this down.
We are running Tomcat 5.5 and are looking to help address any code
that may not be threadsafe. We do use a lot of libraries so not sure
if there is something simple we should be doing like pre-loading all
the libraries or something. We also have MaxThreads set to about 500
, not sure if this is set to high for a JRuby application (4GB Heap on
8GB RAM). Our QA/Capacity team is just getting concerned that every
time we throw load we get massive amounts of these errors in the
Tomcat logs which is raising scaling concerns.
If you have any advice on how to help track the error down in the code
itself (right now we just get a very long backtrace of Jruby) I would
greatly appreciate it. I’ll definitely keep the list posted on my
findings…
Adam
On Wed, Apr 22, 2009 at 7:59 PM, Jay McGaffigan [email protected]
wrote:
Feel free to reach out even if you want to just talk shop. Sometimes
On Wed, Apr 22, 2009 at 11:04 AM, Jay McGaffigan
http://xircles.codehaus.org/manage_email
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email