Xymon Mailing List Archive search

bbwin 0.9 available

3 messages in this thread

list Etienne Grignon · Mon, 18 Dec 2006 01:36:31 +0100 ·
Hello,

First at all, I'm deeply sorry for the large delay between 0.8 and
0.9. I will try to do not disappoint you the next time ...

For the 0.9, these are the news :


+ New Features
* Msgs agent reports events username and you may filter event with user name
* Msgs agent reports some statistics about each event log file. You can activate
  an option which alert you when an event log is full. See the help
chm file for details.
* New BBWinUpdate agent : it used for the semi centralized mode. You
can configure the agent
  to download configurations updates from a central hobbit server.
* First version of BBWin centralized mode implemented. Backend for
hobbit not coded.
  Look at the BBWin.cfg example in the doc folder or in the documentation

+ Updates
* system uptime sometimes returns 0 seconds on the first query. BBWin now send
  a green status when this happens to avoid incorrect alerts.
* msgs now formats description field in status so you don't have to
scroll your window
  to read a large event description
* disk agent reported alerts on unrecognized disks (like disks part of
cluster resources on
  the inactive node of an MSCS cluster). These disks are now
automatically ignored.
* cluster.exe is more verbose when it check the Windows version
* procs agent now is using rules with signs '+' or '-' because XML
format doesn't accept
  '<' sign in attributes ... However, the compatibility with '<' and
'>' is preserved.
* msgs summary option is removed. It is enabled by default and there
is no use to disable it.

+ Bugs
* limit option for cpu agent was wrong in the number of lines
reported, it is fixed.
* When an agent had no configuration in BBWin.cfg and if it was
loaded, BBWin crashed, it is fixed.
* msgs agent had a memory leak when an initialization was made to get
event description. it is fixed.
* wlbs script external had a bug to load the configuration file. it is fixed.
* cpu agent were always missing one process in the ps report. it is fixed.
* externals agent was crashing BBWin if its configuration was empty.
it is fixed.


For the centralized mode, there are no backend on the hobbit part for
the moment ... so, you can enable the bbwin centralized mode it to see
that it is working (You will be able to see the small link
"clientdata" in each status sub page of your host) but it is not
usable at this time.

The others news are that I tested BBWin 0.8 and 0.9 on Windows Vista
Ultimate, and it is working just fine.

I'm using BBWin on over 300 productions servers. The only problems I
got are resolved with the bug fixes included in BBWin 0.9. With the
bbwinupdate agent, you can centralize all or some of your
configuration rules even if it is not the fully hobbit centralized
mode that every body is expecting for.

Here the link for the download :

http://sourceforge.net/projects/bbwin

As always, let me know if you get troubles with this new version.


Regards,

Etienne
list Vernon Everett · Mon, 18 Dec 2006 10:54:37 +0900 ·
Hi all

I detest reinventing wheels, so let's first check if they exist.

Has anybody got anything to monitor Squid and/or Dansguardian?

Regards
    Vernon
list Alun Watson · Tue, 19 Dec 2006 15:06:31 +1300 ·
Hello Vernon
        I've written a script that gathers data from the squidclient
util (just called client in the earlier versions).
I've just started collecting data so I don't have error status's set up
yet ... 

The squidclient util will give can give you ridiculous amounts of data
on squid memory usage, disk IO, 5min and 60min averages etc.
Squidclient mgr:menu will tell you what you can get.

You should be able to change the client script to grab what ever data
you want ...
On the hobbit server

Entered in to hobbitgraph.cfg


[squid]
        TITLE Squid IN-OUT
        YAXIS # bits
        DEF:serverhttpbitsin=squid.rrd:serverhttpbitsin:AVERAGE
        DEF:serverhttpbitsout=squid.rrd:serverhttpbitsout:AVERAGE
        DEF:clienthttpbitsin=squid.rrd:clienthttpbitsin:AVERAGE
        DEF:clienthttpbitsout=squid.rrd:clienthttpbitsout:AVERAGE
        AREA:serverhttpbitsin#@COLOR@:Server Traffic in
        LINE3:serverhttpbitsout#@COLOR@:Server Traffic out
        AREA:clienthttpbitsin#@COLOR@:Client Traffic in
        LINE2:clienthttpbitsout#@COLOR@:Client Traffic out
        COMMENT:\n
        GPRINT:clienthttpbitsin:LAST:Client Traffic in \: %5.1lf%s (cur)
        GPRINT:clienthttpbitsin:MAX: \: %5.1lf%s (max)
        GPRINT:clienthttpbitsin:MIN: \: %5.1lf%s (min)
        GPRINT:clienthttpbitsin:AVERAGE: \: %5.1lf%s (avg)\n
        GPRINT:clienthttpbitsout:LAST:Client Traffic out \: %5.1lf%s
(cur)
        GPRINT:clienthttpbitsout:MAX: \: %5.1lf%s (max)
        GPRINT:clienthttpbitsout:MIN: \: %5.1lf%s (min)
        GPRINT:clienthttpbitsout:AVERAGE: \: %5.1lf%s (avg)\n
        GPRINT:serverhttpbitsin:LAST:Server Traffic in \: %5.1lf%s (cur)
        GPRINT:serverhttpbitsin:MAX: \: %5.1lf%s (max)
        GPRINT:serverhttpbitsin:MIN: \: %5.1lf%s (min)
        GPRINT:serverhttpbitsin:AVERAGE: \: %5.1lf%s (avg)\n
        GPRINT:serverhttpbitsout:LAST:Server Traffic out \: %5.1lf%s
(cur)
        GPRINT:serverhttpbitsout:MAX: \: %5.1lf%s (max)
        GPRINT:serverhttpbitsout:MIN: \: %5.1lf%s (min)
        GPRINT:serverhttpbitsout:AVERAGE: \: %5.1lf%s (avg)\n
[squidclients]
        TITLE Squid Clients
        YAXIS # count
        DEF:numberofclients=squidclients.rrd:numberofclients:AVERAGE
        LINE2:numberofclients#@COLOR@:Number_of_clients
        COMMENT:\n
        GPRINT:numberofclients:LAST:Number_of_clients \: %5.1lf%s (cur)
        GPRINT:numberofclients:MAX: \: %5.1lf%s (max)
        GPRINT:numberofclients:MIN: \: %5.1lf%s (min)
        GPRINT:numberofclients:AVERAGE: \: %5.1lf%s (avg)\n


Entered into hobbitserver.cfg
Added to TEST2RRD="squid=ncv,squidclients=ncv"
Added to GRAPHS="ncv,squid,squidclients"
Create these new lines
NCV_squid="clienthttpbitsin:GAUGE,clienthttpbitsout:GAUGE,serverhttpbits
in:GAUGE,serverhttpbitsout:GAUGE,*:NONE"
NCV_squidclients="numberofclients:GAUGE,*:NONE"


On the hobbit client

Create Squid.sh in the ext directory
#!/bin/sh
#Script to
COLUMN=squid    # Name of the column
COLUMN1=squidclients
COLOR=green             # By default, everything is OK
MSG="Squid Status"
# Get Squid traffic in and out.
VAR2=`/usr/local/squid/bin/squidclient mgr:5min | awk
'/server.http.kbytes_/||/client_http.kbytes_/{print $0}' | sed -e
's/=/:/g' -e 's/\/sec//g' -e 's/server\.http\.kbytes_/serverhttpbits/g'
-e 's/client\_http\.kbytes_/clienthttpbits/g' |awk '{print
$1,$2,$3*8*1024}'`
# Get number of connected clients.
VAR3=`/usr/local/squid/bin/squidclient mgr:info | awk '/Number of
clients accessing cache/{print $0}'| sed -e 's/Number of clients
accessing cache\:/numberofclients\ \:/' -e 's/\//g' -e 's/\ //g'`
# Tell Hobbit about it
$BB $BBDISP "status $MACHINE.$COLUMN $COLOR `date`
<br><b>Data in and out of cache 5 minute average</b>
$VAR2
"
$BB $BBDISP "status $MACHINE.$COLUMN1 $COLOR `date`
$VAR3
"
exit 0

Add entry to run the script in clientlaunch.cfg

[squid]
        DISABLED
        ENVFILE $HOBBITCLIENTHOME/etc/hobbitclient.cfg
        CMD $HOBBITCLIENTHOME/ext/squid.sh
        LOGFILE $HOBBITCLIENTHOME/logs/squid.log
        INTERVAL 5m


That gives me pretty graphs about traffic and number of clients.

Hope that helps.


Alun
quoted from Vernon Everett

-----Original Message-----

From: Everett, Vernon [mailto:user-36f80bd657a9@xymon.invalid]
Sent: Monday, 18 December 2006 2:55
To: user-ae9b8668bcde@xymon.invalid
Subject: [hobbit] Squid & Dansguardian

Hi all

I detest reinventing wheels, so let's first check if they exist.

Has anybody got anything to monitor Squid and/or Dansguardian?

Regards
    Vernon