Hi, My interpretation of the problem is that it may not be related to heap. >From the statistics, the memory heap consumption seems to be within bearable limits. I suspect that the the first servlet after being heavily loaded is starving susbsquent requests for want of some resources critical to its processing. Are you using any resource pools in the servlet that are synchronized? This seems to be the most likely reason since your cpu utilisation is also very low, which points that you may have some I/O wait cycles blocking other requests (or some other synchronous blocking activity).
-- Uddhav ----- Original Message ----- From: "Raja Sekhar" <[EMAIL PROTECTED]> To: "Tomcat Users List" <[EMAIL PROTECTED]> Sent: Friday, March 14, 2003 11:41 AM Subject: RE: Increase heap for a specific webapplication > Hi Shapira, > > I have actually increased the heap size on my server. Its a 4 CPU processor with 4 GB Ram. So I increased my VM options when starting the tomcat to start with 128M and go upto 1.24GB. I suppose thats pretty huge for that. The problem here is, it doesn't max out the memory. When the tomcat stops responding, I check for the heap size and the CPU processes. Both are normal. For example, these are the statistics what I have when my CPU is actually hanging. Just have a look at them > > The total memory occupied by the tomcat was 145 M of which > 121M was resident. This is the top output > > last pid: 5794; load averages: 0.16, 0.19, 0.26 > 20:43:38 > 45 processes: 44 sleeping, 1 on cpu > CPU states: 37.1% idle, 61.5% user, 1.4% kernel, 0.0% iowait, 0.0% swap > Memory: 512M real, 130M free, 299M swap in use, 1099M swap free > > PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND > 5370 gsatpath 21 59 0 130M 108M sleep 27:39 24.82% java > 5521 tomcat 29 59 0 145M 121M sleep 14:28 2.95% java > 5794 gsatpath 1 59 0 2584K 1696K cpu 0:00 0.09% top > 429 root 12 58 0 2464K 2152K sleep 0:56 0.00% mibiisa > 161 root 5 58 0 3216K 2272K sleep 0:27 0.00% automountd > 5496 tomcat 9 59 0 96M 23M sleep 0:04 0.00% java > 321 root 1 30 0 2984K 1640K sleep 0:03 0.00% sshd > 153 root 1 0 0 2016K 1264K sleep 0:00 0.00% lockd > 152 daemon 4 1 0 2520K 1824K sleep 0:00 0.00% statd > 49 root 5 2 0 1344K 848K sleep 0:00 0.00% syseventconfd > 47 root 9 12 0 1536K 1192K sleep 0:00 0.00% syseventd > 366 root 5 23 0 3144K 2216K sleep 0:00 0.00% dmispd > 226 root 5 31 0 2688K 2072K sleep 0:00 0.00% vold > 5276 root 1 38 0 1816K 1448K sleep 0:00 0.00% in.telnetd > 4510 root 1 39 0 1816K 1448K sleep 0:00 0.00% in.telnetd > > Can any one let me know why tomcat can hang in such stage??? > > ..Raj > -- > > On Thu, 13 Mar 2003 12:01:00 > Shapira, Yoav wrote: > > > >Howdy, > >There is one heap per JVM, period. It has nothing to do with tomcat or > >web applications. You can tune this heap extensively using the various > >java runtime options: http://java.sun.com/docs/hotspot/VMOptions.html > > > >If it's an emergency and in production, just allocate more memory to the > >overall heap for now. But you should immediately start to profile your > >applications, identify any leaks, and fix them, all in a development > >environment. > > > >Yoav Shapira > >Millennium ChemInformatics > > > > > >>-----Original Message----- > >>From: Raja Sekhar [mailto:[EMAIL PROTECTED] > >>Sent: Thursday, March 13, 2003 11:18 AM > >>To: [EMAIL PROTECTED] > >>Subject: Increase heap for a specific webapplication > >> > >>Hi, > >> > >>I have two web applications running in my Tomcat 4.0.6. One application > >is > >>data centric and heavily loaded and other is very light with hardly any > >>users hitting. Now what I observe is sometimes if there is a heavy > >search > >>made on the first system... it basically stops responding. I cant send > >any > >>further requests till restart. But if I hit the second webapp, it still > >>keeps responding. > >> > >>So what I feel is tomcat is internally dividing the heap and virtual > >memory > >>for all the webapps equally.. Is there anyway I can alter this heap for > >>each webapp separately?? > >> > >>Or please let me know any other alternative for this problem asap > >because > >>this is a production problem. > >> > >>Thanks & Regards, > >> > >>..Raj > >> > >> > >>Need a new email address that people can remember > >>Check out the new EudoraMail at > >>http://www.eudoramail.com > >> > >>--------------------------------------------------------------------- > >>To unsubscribe, e-mail: [EMAIL PROTECTED] > >>For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > > > >This e-mail, including any attachments, is a confidential business communication, and may contain information that is confidential, proprietary and/or privileged. This e-mail is intended only for the individual(s) to whom it is addressed, and may not be saved, copied, printed, disclosed or used by anyone else. If you are not the(an) intended recipient, please immediately delete this e-mail from your computer system and notify the sender. Thank you. > > > > > >--------------------------------------------------------------------- > >To unsubscribe, e-mail: [EMAIL PROTECTED] > >For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > > Need a new email address that people can remember > Check out the new EudoraMail at > http://www.eudoramail.com > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
