It's funny that a packet with bela in it would cause somebody's program to
crash.
View the original post :
http://www.jboss.com/index.html?module=bbop=viewtopicp=4136855#4136855
Reply to the post :
http://www.jboss.com/index.html?module=bbop=postingmode=replyp=4136855
My immediate thought it is there is some packet being received that has a bad
size value, which is why a 1.6GB buffer might have been allocated.
Maybe you could patch the code that allocates that array to throw an exception
when a certain size limit is reached? I'd also try to run tcpdump and
I'll tell our infrastructure to try and grab that information next time they
get the opportunity which may be a while since it they have no way to grab the
dumps unless the problem happens to freeze the server which is usually just
crashes.
Looking at the stack traces I'm not sure what would
I have some more info. If found the line of code that is allocating the 1.6gb
byte[].
BasicConnectionTable.run():577
| 575: len=in.readInt();
| 576: if(len buf.length)
| 577: buf=new byte[len];
|
So you were correct it
Another update. This bad packet contains the the following
0x62, 0x65, 0x6C, 0x61, 0x00, 0x00, 0x00, 0x00
Which converted to ascii is bela so I think this is pretty obvious that it is
a jgroup/jbossCache originated packet or fragment of a packet
Any other ideas?
View the original post :
Actually I lied the packet's contents are:
[98, 101, 108, 97, 17, 2, 4, 10, 108, 34, 34, 0, 0, -30, -112]
the first 4 bytes == bela
10.108.34.34 is my server's IP address
I don't know what the rest is.
View the original post :
Suggest you post on [EMAIL PROTECTED] instead. This is a detailed JGroups
question, not a JBoss Cache one. Not criticizing you for posting here; you
didn't know that. :-) Just telling you where you're more likely to get a
useful answer.
View the original post :