Looks like you ran out of process memory. How big is your swap file?
Qingtian Wang wrote:
Thanks, I got it in "messages"!!!
-----------
Jan 19 06:39:04 localhost kernel: Out of Memory: Killed process 28274 (java).
-----------
Well, now at least I know it's some memory issue. I wonder whether
this means the OS runs out of memory, or just the Java VM.
Here's what I see in the geronimo console about "Server Info":
--------------
Kernel
Kernel Boot Time Thu Jan 19 20:52:34 CST 2006
Kernel Up Time
3 days 37 minutes 12 seconds
JVM
Java Version 1.4.2_10
Java Vendor Sun Microsystems Inc.
Node localhost/127.0.0.1
Current Memory Used
37.89 MB
Most Memory Used
39.93 MB
Total Memory Allocated
63.56 MB
Available Processors 1
-----------------------
Does this mean, when the memory demanded to run the apps in the
container exceeds the "total allocated", the VM dies by itself, or OS
kills the Java process because memory is out for the entire OS? (I
have 512M ram, and 2G disk swap on this Linux PC.)
Thanks!!!
Qingtian
On 1/22/06, Matt Hogstrom <[EMAIL PROTECTED]> wrote:
/var/log/messages I think unless GC is different.
Qingtian Wang wrote:
Sorry about the ignorance. But where is the syslog? I checked in
/var/log/ but nothing named syslog in there....
Thanks,
Qingtian
On 1/20/06, lichtner <[EMAIL PROTECTED]> wrote:
I didn't catch whether you looked at syslog or not. Anything there?
On Fri, 20 Jan 2006, Qingtian Wang wrote:
This time it ran for more than two days. But I had several occasions
where it died the next day....
Qingtian
On 1/19/06, lichtner <[EMAIL PROTECTED]> wrote:
How long does it run for? Is it always the same amount of time?
On Thu, 19 Jan 2006, Matt Hogstrom wrote:
you could check /var/logs and look in the messages to see if something was
logged at the same time. Other than that others probably have better ideas.
Qingtian Wang wrote:
OK, this time it lived more than one day (more than two days). I found
it dead this morning, and "ps" command shows nothing. The process is
gone.
Looking in geronimo.out and geronimo.log doesn't give much a clue. Is
there any other log I could check? I am not real versed with Linux; is
there some sys log that would show when and how the process got
killed?
Thanks!
Qingtian
On 1/16/06, Paul McMahan <[EMAIL PROTECTED]> wrote:
Intriguing problem.. Could the Geronimo process still be running but is
just inaccessible? You can check by waiting until the server has seemingly
terminated and then type :
ps -ef | grep geronimo
If you see something like :
username 4589 1 11 13:41 pts/2 00:00:39
/usr/java/j2sdk1.4.2_10/bin/java -Djava.endorsed.dirs=
-Dorg.apache.geronimo.base.dir [remainder truncated by me]
then geronimo may have become inaccessible due to, perhaps, an IP address
change (thanks to Leigh Williamson for this suggestion).
As a test of this theory I changed the IP address on my redhat linux system
and was still able to access the admin console. But command line utilities
such as shutdown.sh and deploy.sh no longer worked:
[EMAIL PROTECTED] geronimo-1.0]$ bin/deploy.sh list-modules
Error: Unable to connect to server at deployer:geronimo:jmx --
javax.naming.NameNotFoundException: JMXConnector
Best wishes,
Paul
On 1/16/06, Qingtian Wang <[EMAIL PROTECTED] > wrote:
Hi,
I am running on Fedora as a normal user. Once started, the server runs
fine until the next day, it is stopped with no clear warning in the
geronomo.out log file. How can I keep the server alive all the time.
Do I have to be a root user when starting the server?
Thanks!
Qingtian
