What should I collect for a RV daemon high memory footprint issue?
Article ID: KB0094649
Updated On:
| Products | Versions |
|---|---|
| TIBCO Rendezvous | - |
| Not Applicable | - |
Description
Resolution:
Environment:
= = = = = =
All
Description:
= = = = = =
There are a few factors that could affect RVD/RVRD daemon memory usage.
1). Message load, including protocol messages.
2). Message backlog, usually seen with RVRD across a WAN. Every WAN connection has a maximum capacity. Routing daemons cannot exceed this physical limitation. When the volume of routed data is greater than WAN capacity, RVRD buffers the outbound data.
3). A network problem.
4). A lot concurrent clients.
5 ). A lot of RVRD neighbors.
6). A lot of local networks.
7). Reliability.
8). High retransmission rates.
Resolution:
= = = = = =
Collect the following information during a problem period:
1). Run tibrvlisten on subject “_RV.>” for 45 minutes, using the same network and service parameters as your application is using and send us the output.
2). From the web interface of the problematic RVD, find the number of clients and number of connections. If RVRD, go to the RVRD web interface and make screen captures of page “Services” and “Connected Neighbors”.
3). Details on size and rate of messages being published through the RVD.
4). Monitor the memory and CPU usage for the system and RVD process.
The following is a sample script. It may need to bemodified it to suit your OS.
--------------
#!/bin/sh
while true
do
now=`date "+%m/%d/%y:%H:%M:%S" `
psStatus=` ps -ef -o pid,user,nice,vsz,rss,pcpu,args|grep rvd|grep -v grep` echo "#####################$now#####################"
echo "$psStatus"
top -b
sleep 60
done
-------------
5). Run “netstat -a” and “netstat -s” on the problematic box before and after you run rvtrace/tcpdump and send in the output for review.
6). Run tcpdump/rvtrace on the problematic box, and when the RVD memory usage starts to grow, send in the rvtrace raw packet captures to us for analysis.
rvtrace -w <capture_filename> -i <interface_name> -filter ip
7). When you see the RVD memory starts to grow, enable profiling for the RV daemon for 10 minutes and send in the RVD log file for review.
Environment:
= = = = = =
All
Description:
= = = = = =
There are a few factors that could affect RVD/RVRD daemon memory usage.
1). Message load, including protocol messages.
2). Message backlog, usually seen with RVRD across a WAN. Every WAN connection has a maximum capacity. Routing daemons cannot exceed this physical limitation. When the volume of routed data is greater than WAN capacity, RVRD buffers the outbound data.
3). A network problem.
4). A lot concurrent clients.
5 ). A lot of RVRD neighbors.
6). A lot of local networks.
7). Reliability.
8). High retransmission rates.
Resolution:
= = = = = =
Collect the following information during a problem period:
1). Run tibrvlisten on subject “_RV.>” for 45 minutes, using the same network and service parameters as your application is using and send us the output.
2). From the web interface of the problematic RVD, find the number of clients and number of connections. If RVRD, go to the RVRD web interface and make screen captures of page “Services” and “Connected Neighbors”.
3). Details on size and rate of messages being published through the RVD.
4). Monitor the memory and CPU usage for the system and RVD process.
The following is a sample script. It may need to bemodified it to suit your OS.
--------------
#!/bin/sh
while true
do
now=`date "+%m/%d/%y:%H:%M:%S" `
psStatus=` ps -ef -o pid,user,nice,vsz,rss,pcpu,args|grep rvd|grep -v grep` echo "#####################$now#####################"
echo "$psStatus"
top -b
sleep 60
done
-------------
5). Run “netstat -a” and “netstat -s” on the problematic box before and after you run rvtrace/tcpdump and send in the output for review.
6). Run tcpdump/rvtrace on the problematic box, and when the RVD memory usage starts to grow, send in the rvtrace raw packet captures to us for analysis.
rvtrace -w <capture_filename> -i <interface_name> -filter ip
7). When you see the RVD memory starts to grow, enable profiling for the RV daemon for 10 minutes and send in the RVD log file for review.
Issue/Introduction
What should I collect for a RV daemon high memory footprint issue?
Feedback
Yes
No
Powered by
