Can you run a load test from another computer on the local network?
Something like apache AB gives a histogram and you should be able to
tell what percentage of your requests is the 5 seconds range. It is
best to run it directly to the tomcat private IP that is used inside
the local network, to mak
sult?
>>
>>
>
> I have another production box with different code, but same TomCat
> configuration. And time to time I have the same issues.
>
> TomCat server side code on another box works fast enough (<5 ms per request)
>
> --
> View this message in context:
&g
code on another box works fast enough (<5 ms per request)
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26888782.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
--
onse 300 ms
> (for external users via browser)
>
> BUT time to time request/response time is upto 5 seconds for the same users.
> How I could figure out a problem? How could I check if it is Tomcat server
> problem or not?
>
> You are very welcome with comments.
On 22/12/2009 12:55, Rainer Jung wrote:
On 22.12.2009 13:26, DOrlov wrote:
Rainer Jung-3 wrote:
You can add "%D" to the pattern for the omcat access log, which will
output the response time in milliseconds how Tomcat determined it.
Already done for most cases values are 0,1,2 ms
Max valu
On 22.12.2009 13:26, DOrlov wrote:
Rainer Jung-3 wrote:
You can add "%D" to the pattern for the omcat access log, which will
output the response time in milliseconds how Tomcat determined it.
Already done for most cases values are 0,1,2 ms
Max value which I ever find in the log ~ 20 ms
e log ~ 20 ms
>
>
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26887542.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
-
To unsubscribe, e-mail:
On 22.12.2009 13:13, DOrlov wrote:
Pid Ster wrote:
Looks OK.
When you say the response takes 5s, how are you measuring this?
Is it sustained, or is it occasional?
Is the server heavily loaded at the time - and again how are you
measuring that?
What else could be happening? E.g. sending
During server work we have meta data reloading time frames ~20 sec per hour
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26887423.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
--
On 22.12.2009 12:58, Pid wrote:
On 22/12/2009 11:34, DOrlov wrote:
Pid Ster wrote:
Any chance you can upgrade to the most recent Tomcat?
How much total physical RAM do you have?
Examine the heap "jmap -heap", and post the results.
I have 8Gb RAM on the machine.
PS Old Generation
4
On 22/12/2009 11:34, DOrlov wrote:
Pid Ster wrote:
Any chance you can upgrade to the most recent Tomcat?
How much total physical RAM do you have?
Examine the heap "jmap -heap", and post the results.
I have 8Gb RAM on the machine.
PS Old Generation
47.376815763383185% used
PS Pe
40995788574219MB)
free = 47542480 (45.34004211425781MB)
60.829337266299945% used
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26886984.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
On 22/12/2009 10:18, DOrlov wrote:
Pid Ster wrote:
What about the exact versions of your OS, JVM and Tomcat?
Is your OS a 64bit one?
Correct, I use 64bit machine.
JVM:
Java(TM) SE Runtime Environment (build 1.6.0_17-b04)
Java HotSpot(TM) 64-Bit Server VM (build 14.3-b01, mixed mode)
OS:
-mrsn
Linux 2.6.21.7-2.fc8xen x86_64
TomCat:
apache-tomcat-6.0.14
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26886252.html
Sent from the Tomcat - User mailing list archive a
filename
> to move gc logs into separate file.
>
>
Thank you, in this case I'll re-configure server to support it
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26886161.html
Sent fro
ance problems because of the gc logging (yet).
btw add
-Xloggc: logfilename
to move gc logs into separate file.
regards
Leon
>
> We tried to reproduce issue by collecting production logs and reproducing
> them via tcpreplay to test server, but we weren't able to reproduce issue
&
On 22/12/2009 05:48, DOrlov wrote:
Zacheusz Siedlecki-2 wrote:
This could be problem with memmory and GC. Please show your jvm
memmory settings.
Regards,
Zacheusz
Thank you for reply, I use next memory and GC settings:
-Xms1024m -Xmx6144m -XX:MaxPermSize=512m
27; GC logging?
We tried to reproduce issue by collecting production logs and reproducing
them via tcpreplay to test server, but we weren't able to reproduce issue
:((
Thanks,
Denis.
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26885080.htm
24m -Xmx6144m -XX:MaxPermSize=512m
> -XX:+UseConcMarkSweepGC -XX:+CMSIncrementalMode
>
> --
> View this message in context:
> http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26884294.html
> Sent from the Tomcat - User mailing list archive at Nabble.com.
>
>
>
oncMarkSweepGC -XX:+CMSIncrementalMode
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26884294.html
Sent from the Tomcat - User mailing list archive at Nabble.com.
-
To unsubscr
rver works fine and ~ time for request/response 300 ms
> (for external users via browser)
>
> BUT time to time request/response time is upto 5 seconds for the same users.
> How I could figure out a problem? How could I check if it is Tomcat server
> problem or not?
>
> You are very we
ds for the same users.
How I could figure out a problem? How could I check if it is Tomcat server
problem or not?
You are very welcome with comments.
Thanks,
Denis.
--
View this message in context:
http://old.nabble.com/TomCat-6-performance-issues-tp26884227p26884227.html
Sent from the Tom
22 matches
Mail list logo
