[smokeping-users] Problem with echoping probing a webserver

[Dan Tucny]

Hi,

I've been having some problems with Smokeping whereby it has lots of
gaps in the graphs for all hosts and types of probing... I have spent
quite a while investigating this... I changed smokeping so that it would
keep running in debug mode so that I could see what was happening... 

What I saw was that over a 1hr period of time, when there should have
been 12 runs of echoping, there had been 8... This suggested to me that
for some reason smokeping was getting stuck and not executing any more
checks for a period of time... I looked at process accounting, and this
gave me more direction, it looked like a copy of smokeping & echoping
had been running for around 15 mins before terminating... So, attemping
to reproduce this I sat down and ran echoping by hand over and over, it
did stop at one point, but I ctrl-c'd out, and decided to try the same
again using 'time' ie... 

'time /usr/bin/echoping -h http://server/url -n 20 server'

Eventually it stopped again, this is the output...

Elapsed time: 0.211013 seconds
Elapsed time: 0.160601 seconds
Elapsed time: 0.203720 seconds
Elapsed time: 0.177903 seconds
Elapsed time: 0.147640 seconds
Elapsed time: 1018.960109 seconds
Elapsed time: 0.209551 seconds
Elapsed time: 0.194006 seconds
Elapsed time: 0.158965 seconds
Elapsed time: 0.201483 seconds
Elapsed time: 0.185196 seconds
Elapsed time: 0.144423 seconds
Elapsed time: 0.177843 seconds
Elapsed time: 0.168792 seconds
Elapsed time: 0.234728 seconds
Elapsed time: 0.220822 seconds
Elapsed time: 0.175307 seconds
Elapsed time: 0.172089 seconds
Elapsed time: 0.155299 seconds
Elapsed time: 0.170275 seconds
---
Minimum time: 0.144423 seconds (1773 bytes per sec.)
Maximum time: 1018.960109 seconds (0 bytes per sec.)
Average time: 51.121488 seconds (5 bytes per sec.)
Median  time: 0.177873 seconds (1439 bytes per sec.)

real    17m4.544s
user    0m0.010s
sys     0m0.000s

Wondering if echoping ran without a timeout value if one wasn't
specified, I specified one with '-t 5' on the command line,
unfortunately it didn't help, as can be seen below...

Elapsed time: 0.271899 seconds
Elapsed time: 0.136564 seconds
Elapsed time: 0.160356 seconds
Elapsed time: 0.165629 seconds
Elapsed time: 0.150872 seconds
Elapsed time: 0.146421 seconds
Elapsed time: 0.166235 seconds
Elapsed time: 0.164552 seconds
Elapsed time: 0.191441 seconds
Elapsed time: 0.267351 seconds
Elapsed time: 0.161892 seconds
Elapsed time: 0.155424 seconds
Elapsed time: 0.143079 seconds
Elapsed time: 0.163546 seconds
Elapsed time: 0.144591 seconds
Elapsed time: 0.175590 seconds
Elapsed time: 0.177077 seconds
Elapsed time: 0.191823 seconds
Elapsed time: 0.150354 seconds
---
Warning: 1 message(s) lost (5 %)
Minimum time: 0.136564 seconds (1875 bytes per sec.)
Maximum time: 0.271899 seconds (942 bytes per sec.)
Average time: 0.172878 seconds (1481 bytes per sec.)
Median  time: 0.163546 seconds (1565 bytes per sec.)

real    17m30.010s
user    0m0.000s
sys     0m0.000s

It has actually realised it's lost a request now, but it still took
17mins to get to this point...

As such, I've come to the conclusion that this isn't a smokeping
problem, but an echoping problem, and smokeping just gets caught up in
waiting for echoping to end... 

I suppose this should go to the echoping crew, however I thought I would
bounce it off you guys first, see if anyone else had experienced this
and if so, did you find a way around it?

Thanks

Dan Tucny


--
Unsubscribe mailto:smokeping-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:smokeping-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/smokeping-users
WebAdmin    http://www.ee.ethz.ch/~slist/lsg2.cgi