hi
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
What do I do?
Sameer.
On Friday 03 January 2003 01:29 am, Sameer D. Sahasrabuddhe wrote:
- LUG meet on 12 Jan. 2003 @ VJTI
hi
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
Does "kill -9 pid" also not kill the process?
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
What do I do?
Sameer.
On Fri, Jan 03, 2003 at 01:29:01AM +0530, Sameer D. Sahasrabuddhe wrote:
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
We had kept the server running in this state overnight, with kswapd running at full priority - it's using around 40% CPU time - but no results! I guess it's time to restart the server.
How can I get more info about what happened to the server? The user involved says he just aborted some process with "Ctrl-C" but it refused to go away - it was some really heavy program, related to some project work of his. Is there anything information about the system that I should save before I restart?
Sameer.
What does the "ps ex" says for that process ? Which state it shows ? running,sleeping,zombie,traced ??
try kill -HUP pid or kill -TERM pid , lastly kill -9 pid
Richard Correia Development Manager Unitek Information Systems Ltd richardc@unitek.co.in
----- Original Message ----- From: "Sameer D. Sahasrabuddhe" sameerds@it.iitb.ac.in To: "ILUG-Bom" linuxers@mm.ilug-bom.org.in Sent: Friday, January 03, 2003 9:53 AM Subject: Re: [ILUG-BOM] killing dead processes
- LUG meet on 12 Jan. 2003 @ VJTI
On Fri, Jan 03, 2003 at 01:29:01AM +0530, Sameer D. Sahasrabuddhe wrote:
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
We had kept the server running in this state overnight, with kswapd running at full priority - it's using around 40% CPU time - but no results! I guess it's time to restart the server.
How can I get more info about what happened to the server? The user involved says he just aborted some process with "Ctrl-C" but it refused to go away - it was some really heavy program, related to some project work of his. Is there anything information about the system that I should save before I restart?
Sameer.
MTech Student, Reconfigurable Computing Lab, KReSIT, IIT-Bombay.
You are a bundle of energy, always on the go.
-- _______________________________________________
On Fri, 3 Jan 2003, Sameer D. Sahasrabuddhe wrote:
On Fri, Jan 03, 2003 at 01:29:01AM +0530, Sameer D. Sahasrabuddhe wrote:
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
We had kept the server running in this state overnight, with kswapd running at full priority - it's using around 40% CPU time - but no results! I guess it's time to restart the server.
How can I get more info about what happened to the server? The user involved says he just aborted some process with "Ctrl-C" but it refused to go away - it was some really heavy program, related to some project work of his. Is there anything information about the system that I should save before I restart?
Sameer.
have your tried attaching to the process using strace ? that might give some clue where it's stuck.
-Rajesh
Hello SDS, good-morning.
This seems like those perfect race conditions that the previous <2.2 kernel would provide on machines with LOW amounts of RAM.
Try out what rajesh has said....but generally i rarely have been able to get the OS to a stable state from such a condition. Best would be to just give it a "init 6".
Bye for now.
Trevor
On Fri, 2003-01-03 at 11:20, Rajesh Deo wrote:
- LUG meet on 12 Jan. 2003 @ VJTI
On Fri, 3 Jan 2003, Sameer D. Sahasrabuddhe wrote:
On Fri, Jan 03, 2003 at 01:29:01AM +0530, Sameer D. Sahasrabuddhe wrote:
Our server seems to have run into a problem ... there's a dead process with an RSS of 2GB, and it refuses to die no matter what!
The swap is down to zero, the machine is thrashing, but we can't afford to restart!
We had kept the server running in this state overnight, with kswapd running at full priority - it's using around 40% CPU time - but no results! I guess it's time to restart the server.
How can I get more info about what happened to the server? The user involved says he just aborted some process with "Ctrl-C" but it refused to go away - it was some really heavy program, related to some project work of his. Is there anything information about the system that I should save before I restart?
Sameer.
have your tried attaching to the process using strace ? that might give some clue where it's stuck.
-Rajesh
-- I think my career is ruined!
-- _______________________________________________
-
Amish Munshi -
Rajesh Deo -
Richard Unitek -
Sameer D. Sahasrabuddhe -
Trevor Warren