[hobbit] Problems with writing client data while running several clients

[Rolf Schrittenlocher]

Hi Mathias,

there are no identical entries in bb-hosts. I'm almost sure that the 
client info send is using the correct names, once for the hostname and 
once for each alias, because there is the accurate data in 
hobbit/hobbit/data for both names before and after the period the log 
indicates failure. So it is a temporary problem (it occurred for about 
half an hour, when everything went green again). I have the impression 
that if two hosts are sending data at almost the same time hobbit 
doesn't like that. If they do so for a period>status bbdisplay goes 
purple. Once the point of time those two are sending differs for at 
least one second (if my interpretation of the message in the log is 
right) data is stored and everything goes back green. My proposition is 
to make the server wait for 1 second before writing if he can't instead 
of writing the information to a log but perhaps this is infeasable.

kind regards
Rolf 

>Hi Rolf,
>
>have you looked into your hobbit-log files?
>New with hobbit i had the same problem and found out
>that this happens when to identical (in my case conn-tests)
>run on the server and collect the information for the
>same machine:
>
>page core-router
>1.2.3.4  router1 #
>[..]
>page locationA
>1.2.3.4  router1 #
>[..]
>Maybe you find identical lines testing the same in your bb-hosts.
>
>Are the service you are monitoring for the virtual hosts equaly named?
>
>Best
>  Mathias
>
>In message <43CE350C.7020404 at rz.uni-frankfurt.de>, Rolf Schrittenlocher writes:
>  
>
>>Hi,
>>
>>we are running 1 to 3 clients on each host, one for $hostname and one 
>>for each virtual host defined on the machine in order to get client data 
>>like disk, cpu, etc. for $hostname and for each virtual host. Each 
>>client is started using
>>$HOBBITCLIENTHOME/bin/hobbitlaunch 
>>--config=$HOBBITCLIENTHOME/etc/clientlaunch.${MACHINE}.cfg 
>>--log=$HOBBITCLIENTHOME/logs/clientlaunch.${MACHINE}.log 
>>--pidfile=$HOBBITCLIENTHOME/logs/clientlaunch.${MACHINE}.pid
>>
>>Now I have the problem that client data of one or both clients is 
>>neglected by the server. In rrd-data.log it shows:
>>2006-01-18 13:03:40 RRD error updating 
>>/pica/ffm/com/sys/hobbit/hobbit/data/rrd/lbscgi2/netstat.rrd from 
>>141.2.164.215: illegal attempt to update using time 1137585820 when last 
>>update time is 1137585820 (minimum one second step)
>>In rrd-status.log are similar entries like
>>2006-01-18 12:48:38 RRD error updating 
>>/pica/ffm/com/sys/hobbit/hobbit/data/rrd/lbscgi2/memory.real.rrd from 
>>141.2.164.215: illegal attempt to update using time 1137584918 when last 
>>update time is 1137584918 (minimum one second step)
>>
>>The client data (disk, etc) for both becomes purple in bb.html.
>>Date for both clients is stored (sometimes) correctly:
>>hobbit/hobbit/data/rrd/otter
>>-rw-r--r--   1 bb       prod       19576 Jan 18 12:43 memory.real.rrd
>>hobbit/hobbit/data/rrd/lbscgi2
>>-rw-r--r--   1 bb       prod       19576 Jan 18 12:58 memory.real.rrd
>>
>>I tried starting the clients with some seconds difference, that helps 
>>for a certain time and then the problem reappears. What can I do?
>>
>>kind regards
>>Rolf
>>
>>-- 
>>Mit freundlichen Gruessen
>>Rolf Schrittenlocher
>>
>>HRZ/BDV, Senckenberganlage 31, 60054 Frankfurt 
>>Tel: (49) 69 - 798 28908   Fax: (49) 69 798 2881
>>LBS: lbs-f at mlist.uni-frankfurt.de
>>Persoenlich: schritte at rz.uni-frankfurt.de
>>
>>
>>
>>To unsubscribe from the hobbit list, send an e-mail to
>>hobbit-unsubscribe at hswn.dk
>>
>>
>>
>>    
>>
>
>To unsubscribe from the hobbit list, send an e-mail to
>hobbit-unsubscribe at hswn.dk
>
>
>
>  
>


-- 
Mit freundlichen Gruessen
Rolf Schrittenlocher

HRZ/BDV, Senckenberganlage 31, 60054 Frankfurt 
Tel: (49) 69 - 798 28908   Fax: (49) 69 798 2881
LBS: lbs-f at mlist.uni-frankfurt.de
Persoenlich: schritte at rz.uni-frankfurt.de