[Spread-users] very big lag when one spread daemon stopped

Csaba Endre Simon secsaba at gmail.com
Thu May 31 07:30:43 EDT 2007


Hi all,

I am running the Spread daemon v4.0 on two different computer: Linux
(10.1.0.44) and Solaris (10.1.0.51). On the Linux box I am running the
TestSpreadClient and the TestSpreadSvr application:

$ ./TestSpreadClient 1000
$ ./TestSpreadSvr 1000 200 5000

Here is the code
http://commedia.cnds.jhu.edu/pipermail/spread-users/2004-November/002309.html

When killing the daemon on the Solaris box (ctrl-c) there is a very
big lag receiving the message:

...
25: 1
26: 1
27: 1
28: 1
29: 11802
30: 11599
31: 11396
32: 11193
33: 10989
...
83: 828
84: 625
85: 421
86: 218
87: 14
88: 1
89: 1

As you see the daemon was stopped after sending message #28. Between
#29-87 to receive one message takes a very big time. After some time
(message #88) receiving messages come back to normal.

My spread.conf is the following:

Spread_Segment  10.1.0.255:4803 {
    linux   10.1.0.44
    solaris 10.1.0.51
}

What can be the error? What setting do I need to tune this time period?

Best Regards,
Csaba




More information about the Spread-users mailing list