tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Smith <d...@cornell.edu>
Subject Re: very slow class loading on initial JSP/servlet request after restart
Date Fri, 20 Feb 2009 18:36:37 GMT
Sam Hokin wrote:
> Christopher Schultz wrote:
>>
>> jstack.8.txt is the last thread dump where File.exists was stalled.
>>
>>> http://ims.net/jstack/jstack.9.txt
>>> http://ims.net/jstack/jstack.10.txt
>>> http://ims.net/jstack/jstack.11.txt
>>
>> The server appears to be idle, here.
>>
>> It's a little weird that thread 9770 has NO STACK INFO AT ALL.
>>
>>> http://ims.net/jstack/jstack.12.txt
>>> http://ims.net/jstack/jstack.13.txt
>>
>> Also idle.
>
> Yeah, I _thought_ it was still stalled on 9-11, but it looks like I
> misjudged the timing by a bit.  I was flipping back and forth between
> virtual windows, hitting jstack while checking if my browser had the
> response from the server.
>
>> Obviously, the File.exists method shouldn't be taking that long... it's
>> a pretty simple operation. Are you using an NFS or other network share?
>> Does your disk have any physical problems? Is this machine running next
>> to any equipment that generates a lot of stray alpha particles? :)
>
> Not that I know of. :) We now know it's hanging on
> java.io.File.exists(), though, so I suppose that's leading us
> somewhere.  It's a logical RAID array on four physical disks.  I've
> run fsck on it (when this all first happened) and it's fine.  And this
> problem persists no matter where the classes are on the disk; whether
> they're in a JAR file anywhere, or split out under WEB-INF/classes.  I
> don't think it can be a disk issue.  I guess I could write a
> standalone Java routine to play with the File.exists() method and see
> if I can reproduce the delay "manually"....
>
>
It _could_ be a disk issue.  I've seen disks that scan fine contain a
growing number of blocks causing read errors and continuous retries.

--David


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Mime
View raw message