bigsister-devel Mailing List for Big Sister

Has anyone out there looked at using SNMP to monitor memory usage in the Sun
Java VM, using the built-in SNMP agent in the VM?

 

Figured I should ask, before I dive into this myself. ;-)

 

I’d appreciate it if any replies to this post include the cc: to my work
e-mail. (I’m subscribed to the lists from a private e-mail address.)

Peter Værlien | IT Systems Engineer
Tel:  +47 922 58 200 Cell:  +47 917 69 384

Verdande Technology AS
Stiklestadveien 1
N-7041 Trondheim, Norway
 <http://www.verdandetechnology.com> verdandetechnology.com

 

----------------------------------------------------------------------------
--------------------------------------- 

The information contained in this message may be CONFIDENTIAL and is
intended for 
the addressee only. If you have received this communication in error, be
aware that 
forwarding it, copying it or in any way disclosing its content to any other
person, is strictly 
prohibited. Any unauthorized use, dissemination of the information or
copying of this 
message is prohibited. If you are not the addressee, please notify the
sender immediately by 
return e-mail and delete this message. Thank you.

 

 

 

 

Hi,

i already posted this problem a few months ago. Thomas Aeby told me, to use
the keyword "pernode" in the report section. This was quite ok for the check
I was writing at this time, but now I have this problem once again. In
contrast to my first check, this check uses some values from uxmon-asroot to
help identify which task failed (The check monitors directorys which should
be created by backup). If i use "pernode" the values from uxmon-asroot are
not displayed in the webinterface. Leaving the report section as it is
(without pernode) results in correct display but status is displayed twice,
e.g.:

Rbackup for Customer (system) has failed
Rbackup for Customer (system) has failed

I tried to use the check for more than one customer. This leads to the
result, that the last result is used for backup, e.g. I the first backup is
ok, but the second has a fault, Big Sister assumes status red for all
backups. I think the reason for all those problems are the cfg file. I red
the Devel Document carefully, but this things are not really covered in it.

Another failure (Ok, this is more related to beauty) is, that no Status
Lights are displayed left to the status message.

I hope someone has the reason for this. My work lacks due to this since a
few weeks.

Best Regards

Stefan Steimann
===============================================================================

rbackup.cfg

test rbackup{

        static set name "rbackup";

        static set features unix local;

        static set description "Check for possible error Files in Backup
Directorys";

        pernode set report_item "rbu";

        static set arguments

                "dir:string:Directory where Backup is saved"

                "customer:string:Name of customer to which the backup
belongs"

                "machine:string:System which has been backed up"

                "perf:int:report perf data every perf minutes"

                "item:string:Item as which to report.";


        init {

                instance import dir;

                instance import customer;

                instance import machine;

                instance import item report_item;

                instance import perf perf_frequency;

        }

        pernode  monitor {
                get rbackup.hrBackupStat[0];

                pernode set backupstat1 ${rbackup.hrBackupStat[0]};

        }

        report {

                status red  ${backupstat1} == 2  "&red" "Rbackup for "
${customer}"("${machine}")" " has failed";
                status yellow ${backupstat1} == 3 "&yellow" "Rbackup for "
${customer}"("${machine}")""  has unknown failure";
                status green 1 "&green" "Rbackup for "
${customer}"("${machine}")" " has completed Successfully";
              }
        perf {
              export always single.hrBackupStat, ${backupstat1};
             }


}

=================================================================================================

rbackup.pm:


package Requester::rbackup;



use lib "/usr/local/bigsister/uxmon";
use lib "/usr/local/bigsister/bin";


use strict;

use Requester::Requester;



@Requester::rbackup::ISA = qw(Requester::Requester);



my $rbackup = new Requester::rbackup();



sub new

{

   my( $this, $args ) = @_;

   my $class = ref($this) || $this;

   my $self = {};

   bless( $self, $class );

   $self->register("rbackup");

   return( $self );

}



sub getinfo

{

    my( $self ) = @_;

    my $info = {};

    my $budir = $self->{"env"}->expand_value( "dir" );

    my ($Sekunden, $Minuten, $Stunden, $Monatstag, $Monat,

    $Jahr, $Wochentag, $Jahrestag, $Sommerzeit) = localtime(time);

    $Monat+=1;

    $Jahr+=1900;

    my $zeit = "$Stunden$Minuten";

    my $verzeichnis1;
    my $verzeichnis2;

    if ($Monatstag < 10)
    {
    $verzeichnis1 = "$Jahr-0$Monat-0$Monatstag";
    $Monatstag-=1;
    $verzeichnis2 = "$Jahr-0$Monat-0$Monatstag";
    }
    else
    {
    $verzeichnis1 = "$Jahr-0$Monat-$Monatstag";
    $Monatstag-=1;
    $verzeichnis2 = "$Jahr-0$Monat-$Monatstag";
    }

    my $status;

    if ($zeit < "2359" && $zeit > "2000")
    {

        if (-e "$budir/$verzeichnis1/daten.err")

        {

            $status = 2;

        }

        else

        {

            if (-e "$budir/$verzeichnis1/daten.ok")

            {

              $status = 1;



            }

            else

            {

              $status = 3;

            }

        }

    }

    else

    {

        if (-e "$budir/$verzeichnis2/daten.err")

        {

          $status = 2;

        }

        else

        {

            if (-e "$budir/$verzeichnis2/daten.ok")

            {

              $status = 1;

            }

            else

            {

              $status = 3;

            }

        }

    }

    return {0 => { "hrBackupStat" => $status }};
}

1;

Hello all,=20

=20

I=B4m using BigSister  1.02 and I have a few questions I hope BigSister =
gurus can answer or redirect me to the appropriate source:

=20

1)       I=B4m currently sending traps from BigSister to OpenView which =
works fine. Is it possible that BigSister receive traps from OpenView =
(or other monitoring agents)? How can be done?

=20

2)       Currently BigSister uses SNMP Ver. 2c, can be configured in =
order to use  SNMP Ver 3?

=20

3)       As has been reported by other users, alarm reminders show the =
state that an alarm had when raised, however this state may have changed =
 (from yellow to red or vice verse); I think that would be nice that =
reminders reflect  the current state of the alarm. If this is not a bad =
idea, which modifications would be necessary?

=20

=20

Thanks a lot.

=20

Carlos

what causes that cfg to be loaded?

I have a uxmon config that declarse the "cpuload" test on localhost. So I
was expecting up/down monitoring for the cpu load. But I got a down alarm
about the cpu "only 4% idle!" which I can only track down as being an alert
text within CPUperf.cfg.

More importantly, the output from the top command included with the alarm
message shows the cpu was 99.8% idle when the alarm was raised. And the top
output shows it was 98.6% (yes, actually less idle) when the alarm was
cleared. So I was trying to figure out what mis-parsed the output from
top...

The details included with the alarm (the down) are:

----------BEGIN-----------
&red  only 4% idle!
CPU load = 0.31, &green Swap Space 1.2GB (100%) free
top - 04:27:30 up 25 days, 11:42,  0 users,  load average: 0.31, 0.57, 0.30
Tasks:  41 total,   1 running,  40 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.0% us,  0.0% sy,  0.0% ni, 99.7% id,  0.3% wa,  0.0% hi,  0.0% si
Mem:    385896k total,   321200k used,    64696k free,    17608k buffers
Swap:  1269112k total,      144k used,  1268968k free,   239880k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

    1 root      16   0  2464  628  540 S  0.0  0.2   0:00.58 init

--------------END----------
(rest of the proc table snipped)

thoughts?
-c

Escalation CAN be done out-of-the-box!

I'm answering my own question for the list archive in the hope that this
will save someone time some day.

--GOAL:

When an alarm first occurs, notify someone (eg, send an email message).
After some period of time, if the alarm is still active, notify
*additional* people. (As in "Bob didn't fix it after a half hour, better
notify Mary too!")

--SOLUTION:

A little bit of Perl eval() gymnastics in PAGER{} rules within your
bb_event_generator.cfg file. We'll make the first notification of an alarm
go to one recipient (or a set of recipients.) Then we'll make subsequent
*reminders* for the alarm go to more recipients (a light weight sort of
escalation).

The trick is to use:

  (time()-$time)

in your PAGER{} preconditions. That's it, that's all the magic.

time() of course eval()'s to the current wallclock-seconds, and $time evals
to the wallclock-seconds when the alarm was first raised. Aside: in my
first post, I pined for an $age variable in the preconditions evaluation.
An $age variable could easily be defined in
share/bigsister/bin/bs_evgen.pm, but this eval() hack works without having
to edit any code.

--EXAMPLE:

# there should be a TAB after the *.* (all hosts, all tests)
# and bear in mind that this rule will generate *reminders* of
# this alarm every 10 minutes (repeate=10) unless some other rule
# changes that.
*.*    delay=0 repeat=10 mail=my_escalation_group

# Now use PAGER{} to direct alerts that are to "my_escalation_group"
# to where they need to go based on how long this alarm has been
# raised.
#
# Note that the (time()-$time) yields units of seconds, while the
# repeat=10 equate is in minutes. So <=1800 is less-than-or-equal
# 30min. So I'm using <=1850 (50 seconds of elbow room) to make
# sure the initial alert, and the 10min, 20min and 30min reminders
# go to one set of mail= recipients. And the >1850 makes the
# 40min, etc reminders go to the second set of recipients.
#
# Notice that the initial (<30min) recipients are also in the
# second (>30min) set so they get the "up" message too when
# the alarm clears.
#
# You could also add more preconditions; eg, use weekday and daytime
# preconditions to make schedules if you have different people on
# call at different times/days.

PAGER{ $mail eq 'my_escalation_group' and (time()-$time)<=1850 }
mai...@do...
PAGER{ $mail eq 'my_escalation_group' and (time()-$time)>1850 }
mai...@do...,us...@do...

--PITFALLS:

If you're going to use my idea of "fake" recipients, and then expand them
in PAGER{} rules, be *certain* you have a matching PAGER{} rule for all
wallclock times. If you initial recipient "set" is just one email address,
just set that in the mail= equate on the alarm rule. Then user a PAGER{}
precondition with timeframe to add more recipients by setting the mail=
equate to fir...@do...,add...@do... . The "fake" recipients just
gives you fine grain control to break out handling of schedules and such.

Hi all,

I see there's been a long discussion (albeit infrequently) of implementing
escalation...

Is an alarm passed "thru" bb_event_generator.cfg parsing each time it
generates alerts? What if we had a new equate that could be tested in the
PAGER{} rules. It would go a long way to solving escalation if I could just
say...


*.* delay=0 [etc etc] mail=foobar
PAGER{$mail = "foobar" and $age < 1800}  mai...@us...
PAGER{$mail = "foobar" and $age >= 1800} mai...@us...,tw...@us...

I just want to send to more/different destinations as the age of the alarm
increases. I don't need new/additional alarms (eg, I don't need extra
alarms as the host.test changes states grn->yel->red.)

thoughts?
-craig

On Mon, Mar 12, 2007 at 03:22:03PM +0100, stefan steimann wrote:
> I had already posted my problem a few day ago. Niels Baggesen mentioned that
> it?s caused by wrong variable declaration. Well, this was quite right, but
> it?s still not ok. I think that zpool.cfg isn?t able to fetch data from
> zpool.pm.

This not anything that I have tried at all (I maintly do SNMP based
tests), but looking at other tests I think that you will need some
additional packaging to make it work. Unfortunatelt Thomas didn't
include details of using perl based requesters in his tester sequel from
a few christmases back.

If you look at for instance the nut requester, you will it as a perl
package and I guess that is the magic incantiation that you will need:

> package Requester::nut;
> 
> use Requester::SlowRequester;
> 
> my $oldproto_warned = {};
> 
> our( @ISA );
> @ISA = qw(Requester::SlowRequester);
> 
> my $requester = new Requester::nut();
> 
> sub new {
     
etc...

/Niels

-- 
Niels Baggesen - @home - Århus - Denmark - nb...@us...
The purpose of computing is insight, not numbers   ---   R W Hamming

I believe this thread belongs on the General list, so I am cross-posting  
it to there.

The most obvious way to do this is to use the "command" check. Check  
appendix III of the manual ("Old-style Health Checks").

Another check you may want to look at is the "statusfile" check. It is  
also an Old-style check.

Peter

On Mon, 19 Mar 2007 08:38:23 +0100, Alexander Kushnirenko  
<ku...@un...> wrote:

> Hi,
>
> Is there a module to check shell script error code.  I would think of
> something like:
> errcode = 0 - green
> errcode = 1 - yellow
> errcode = 2 - red
>
> I think it man be useful addition to create quick checks for people who
> are not very familiar with perl or with the steps necessary to create
> bigsister module.
>
> Sasha.
>
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share  
> your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Bigsister-devel mailing list
> Big...@li...
> https://lists.sourceforge.net/lists/listinfo/bigsister-devel
>
>



-- 
Peter Værlien
Fritz Aabakkens vei 17, 7072 Heimdal, Norway
Telephone: 917 69 384, E-Mail: pva...@on...,  
http://varlien.home.online.no
It always takes a Clinton to clean up after a Bush - Hillary 2008

On Thu, 2007-03-15 at 13:04 +0300, Alexander Kushnirenko wrote:
> But before we jump into this activity I would like to ask developers is
> it worth doing (may be there are other means to solve this problem)

It's probably worth doing ... actually it's not going to take much 
effort.

>   Or
> perhaps icmp_ping is being replaced by a different module.

Yes, the icmp_ping module is eventually going to be replaced by one that
is able to do concurrent alive checks ... but this is not going to
happen very soon.

Best regards,
Tom
-- 
----------------------------------------------------------------------------
Thomas Aeby, Kirchweg 52, 1735 Giffers, Switzerland, Tel: (+41)264180040
Internet: ae...@gr...                       PGP public key available
----------------------------------------------------------------------------

Hi,

Is there a module to check shell script error code.  I would think of
something like:
errcode = 0 - green
errcode = 1 - yellow
errcode = 2 - red

I think it man be useful addition to create quick checks for people who
are not very familiar with perl or with the steps necessary to create
bigsister module.

Sasha.

On Sun, 18 Mar 2007 15:08:25 +0100, Alexander Kushnirenko  
<ku...@un...> wrote:

> Hi,
>
> We seem to have problem for icmp_ping working under bigsis account
> (debian bigsiter-1.02).  It can nnot open RAW socket.  When you start
> uxmon under root everything is fine.  Any comments on that?

Fact of life. Define your icmp_ping tests in the uxmon-asroot file,  
instead of uxmon-net. That's what it's there for.

;-)

To make this clearer: the presence of a file named uxmon-asroot, which has  
the same syntax as usmon-net, causes an instance of uxmon to be started  
the runs as root. Tests, such as icmp_ping, that require root priviledges  
to work go in uxmon-asroot. All other tests should be put into uxmon-net.

Peter
>
> Sasha.
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share  
> your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Bigsister-devel mailing list
> Big...@li...
> https://lists.sourceforge.net/lists/listinfo/bigsister-devel
>
>



-- 
Peter Værlien
Fritz Aabakkens vei 17, 7072 Heimdal, Norway
Telephone: 917 69 384, E-Mail: pva...@on...,  
http://varlien.home.online.no
It always takes a Clinton to clean up after a Bush - Hillary 2008

Hi,

We seem to have problem for icmp_ping working under bigsis account
(debian bigsiter-1.02).  It can nnot open RAW socket.  When you start
uxmon under root everything is fine.  Any comments on that?

Sasha.

Hi,

We use icmp_ping to test a dozen of connections that are necessary for
smooth operation of university network.  But icmp_ping have only a few
knobs that you can tune.  Most of performace parameters are hardcoded. 
So we would like to add certain parameters:

persantage of lost packets for red alarm
persantage of lost packets fot the warning

RTT for red alarm
RTT for warning.

We have certain lines (128k) which produce tons of alarms.

But before we jump into this activity I would like to ask developers is
it worth doing (may be there are other means to solve this problem)  Or
perhaps icmp_ping is being replaced by a different module.

Sasha.

Hi,

I think it is useful to add the following info about realhttp module in
its documatation:

"To operate the module requires libwww-perl to be installed."

It took me some time to dig out this problem. Hope it can save time to
others.

Sasha.

Hello,

I had already posted my problem a few day ago. Niels Baggesen mentioned tha=
t
it=B4s caused by wrong variable declaration. Well, this was quite right, bu=
t
it=B4s still not ok. I think that zpool.cfg isn=B4t able to fetch data from
zpool.pm.
Are there some special requirements in the Perl Module, so that the *.cfg
can fetch data.
This is my ZFS Module:

zpooldisk.cfg

test zpooldisk
{
        static set name "zpooldisk";
        static set features "local" "unix";
        static set description "ZFS Pool Disk Anzeige";
        static set arguments
                "perf:int:report perf data every perf minutes"
                "item:string:Item as which to report";
        pernode set report_item "zpool";

        pernode monitor
        {
                get zpooldisk.diskstat;
                pernode set dlist ${zpooldisk.diskstat};
        }

        init
        {
                instance import item report_item;
                instance import perf perf_frequency;

        }
        pernode report
        {
                 comment "html" "<hr /><strong>" "Disklist:" "</strong><br>=
"
${dlist}  ;
        }
}

zpooldisk.pm

#!/usr/bin/perl
open (ZPOOL,"/usr/sbin/zpool status |");
my @dstat=3D<ZPOOL>;
my $diskstat =3D join ("", @dstat);

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

I get an output when using a print statement in zpool.pm, but the disklist
isn=B4t printed on webinterface. The last idea I have is that
zpooldisk.cfgisn=B4t able to fetch status due to variable permissions.
I=B4m not very
familiar with perl at the moment, so I don=B4t know exactly about it.

Best Regards

Stefan Steimann

On Mon, Mar 05, 2007 at 03:11:06PM +0100, stefan steimann wrote:
>                pernode set dlist${zpool.disklist};
...
> $diskstat = join("", @ZP2);

disklist <> diskstat

/Niels

-- 
Niels Baggesen - @home - Århus - Denmark - nb...@us...
The purpose of computing is insight, not numbers   ---   R W Hamming

On Mon, 2007-03-05 at 15:11 +0100, stefan steimann wrote:
> I´m trying to develop a check for Solaris zpools. The check seems to
> be ok, but the status is listed twice on the webinterface.

I do not see a corresponding problem in the test description. What does
actually the status look like? Is the whole entry (lamp + text) duped?
Is the zpool test listed only once in uxmon-net?

>  I also tried to display a disklist on the webinterface, but nothing
> is displayed.

That one is easy:
                
>   comment ${dlist}; 

... should read ...


    comment text ${dlist};

Perhaps, you'll want to declare the "report" section "pernode", too.

Best regards,
Tom
-- 
----------------------------------------------------------------------------
Thomas Aeby, Kirchweg 52, 1735 Giffers, Switzerland, Tel: (+41)264180040
Internet: ae...@gr...                       PGP public key available
----------------------------------------------------------------------------

Hello,

I=B4m trying to develop a check for Solaris zpools. The check seems to be o=
k,
but the status is listed twice on the webinterface. I also tried to display
a disklist on the webinterface, but nothing is displayed. I checked the
output of the perl requester, but it seems to be allright. Does anyone know
the mistakes?

This is the check:

zpool.cfg:


        static set name "zpool";
        static set features "local" "unix";
        static set description "Test zum ueberwachen eines ZFS Pools";
        static set arguments
                "perf:int:report perf data every perf mintues"
                "item:string:Item as which to report"
                "status:string:Test goes red if Status is not optimal";
        pernode set report_item "zpool";

        pernode monitor
        {
                get zpool.status;
                pernode set statvar${zpool.status};

                get zpool.disklist;
                pernode set dlist${zpool.disklist};
        }

        init
        {
                instance import item report_item;
                instance import perf perf_frequency;

        }
        report
        {
                 status red (1 > ${statvar} ) "&red"
                "Warning! Status of zpool is not optimal";
                status green 1 "&green"
                "Status of zpool is optimal";
                comment ${dlist};


        }

        perf
        {
                export always single.zpoolstatvar, ${statvar};
        }
}

zpool.pm

#!/usr/bin/perl

$zpcom =3D "/usr/sbin/zpool list |" ;
#print $zpcom;
open IN, "$zpcom" ;

@ ZP =3D <IN> ;
chomp @ZP;
if (@ZP =3D grep /ONLINE/, $ZP[1])
{
$status =3D 1;
#print "$status \n";
}
else
{
$status =3D 0;
#print "Status BAD \n";
}

$zpcom2 =3D "/usr/sbin/zpool status |" ;
open IN2, "$zpcom2" ;

@ZP2 =3D <IN2> ;
close IN2;
$diskstat =3D join("", @ZP2);
#print $diskstat;




=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D

Best Regards

Stefan Steimann

Hello,

On Mon, 2006-07-10 at 17:24 -0500, Hern=C3=A1ndez L=C3=B3pez Carlos wrote=
:
> I have developed a module for monitoring sybase databases. It works
> fine,

Congrats! :-)

>  but  I would like to include some html tags in order to speed up the
> search of information  as
[...]
>       set text =3D=E2=80=9D<A HREF=3D\=E2=80=9D#${database}\=E2=80=9D> =
${label} </A>=E2=80=9D;
>=20
>       status red ( ${status} =3D=3D 2) "&red"  ${text} ;

Ok, I do not want this to work. Why? Status texts are not only appearing
on the status page, but are also used e.g. for alarming. Allowing what
you try to do would imply (besides other implications):

  - every alarming script had to include a HTML parser
  - there were information in the status texts that are useless
    for those not reading them via the status page

If I understand well, in your case the "comment html" sections are
rather lengthy and it's not easy to find the section corresponding to
a bad status. I haven't got an ideal solution for this, but might it
help to add a contents section at the beginning of your html comment?

I assume this should be possible via a "pernode report" such as

  pernode report {
     comment html "<h3>databases</h3> ..."
  }

"pernode" methods are usually executed before the first instance method
of the same host and name is executed, so html comments of a "pernode
report" should appear before those of a normal "pernode".

Of course, you have to assemble the entries for each database, either by
using a "for" or by storing them in a variable in the "monitor" or
"discover" method for later output by "pernode report".

Is this an acceptable approach?

Best regards,
Tom
--=20
-------------------------------------------------------------------------=
---
Thomas Aeby, Kirchweg 52, 1735 Giffers, Switzerland, Tel: (+41)264180040
Internet: ae...@gr...                       PGP public key available
-------------------------------------------------------------------------=
---

Hello all,=20

=20

I have developed a module for monitoring sybase databases. It works =
fine, but  I would like to include some html tags in order to speed up =
the  search of information  as, in some cases, I=B4m monitoring  20 or =
more databases by server.

=20

The current state is :=20

=20

    init {

         ....

        instance set label "${database}"@"${server}";

        ....

    }

=20

    report {

      status red ( ${status} =3D=3D 2) "&red" ${label};

      status yellow ( ${status} =3D=3D 1) "&yellow" ${label};

      status green ( ${status} =3D=3D 0)  "&green" ${label};

      comment "html" "<hr> ${message}";

    }

=20

=20

And I=B4d like to have something like:

=20

    report {

      set text =3D"<A HREF=3D\"#${database}\"> ${label} </A>";

      status red ( ${status} =3D=3D 2) "&red"  ${text} ;

      status yellow ( ${status} =3D=3D 1) "&yellow" ${text};

      status green ( ${status} =3D=3D 0)  "&green" ${text};

      comment "html" "<hr>  <A NAME=3D\"${database}\"> </A> ${message}";

    }

=20

which does not work, because bigsister server takes ${text}  as plain =
text.

=20

Suggestions are appreciated.

=20

Thanks a lot.

=20

Carlos

=20

=20

Hello,

I propose a minor change to the BigSister configuration language:

diff -u big-sister-1.02/bsmonitor/Statusmon/DisplayCFG.pm
/data/bigsis-test/bin/Statusmon/DisplayCFG.pm
--- big-sister-1.02/bsmonitor/Statusmon/DisplayCFG.pm   2005-10-13
15:15:27.000000000 +0200
+++ /data/bigsis-test/bin/Statusmon/DisplayCFG.pm       2006-06-30
16:14:24.147125168 +0200
@@ -275,7 +275,7 @@
                else {
                    $args{"refto"} = $args[0];
                    unless( $args{"refto"} =~ /\.htm/ ) { $args{"refto"} .
= ".html"; };
-                   $args{"refto"} .= "#";
+                   unless( $args{"refto"} =~ /=$/ ) {$args{"refto"} .= "
#";};
                }
            }
            elsif( ($cmd eq "skin") || ($cmd eq "Logskin") ) {

This permits to use things like "%refto notes/hostinfo.html?server=" in
bb-display.cfg, where hostinfo.html can now trigger execution of a cgi
script that is provided with the hostname as a parameter.
I don't expect this change to break anything.

Another minor change:

 diff -u big-sister-1.02/modules/dbstore/bin/DBCapsulator/DBI.pm
/data/bigsis-test/bin/DBCapsulator/DBI.pm
--- big-sister-1.02/modules/dbstore/bin/DBCapsulator/DBI.pm     2006-01-17
14:46:35.000000000 +0100
+++ /data/bigsis-test/bin/DBCapsulator/DBI.pm   2006-07-07
19:26:43.802112751 +0200
@@ -19,6 +19,7 @@
     $self->dbh( $dbh );
     my $table = $self->dbcfg()->{"table"};
     my @names = $dbh->tables( undef, undef, $table, undef );
+    grep ( y/"//d, @names );
     my @names = grep( /^(|.*\W)$table(\W.*|)$/, @names );
     unless( @names ) {
         $self->create_table() || die "table $table does not exist in
$dbcfg->{'datasource'} and create table did not succeed";

This helps to use DB2 for the performance data store. In DB2, table names
are made up of two parts, the "Schema" name and the table name itself. The
Schema name, if not specified, defaults to the current user name. This is
mostly fine, but when the bshistgraph executes in the web server context,
the user is, e.g., "www" while the tables were created by the user
"bigsister". So, in db.cfg, I specify full table names, e.g.
BIGSISTER.BSPERF. But DBI->tables() sadly returns this as
"BIGSISTER"."BSPERF" including the quotes. My patch strips off these
quotes, and everything works as intended. I am not sure, however, that this
does not break other database connectors.


Regards,

Peter Hettkamp

On Tue, 2006-06-13 at 21:27 +0200, Niels Baggesen wrote:
> +LANG=C LC_MESSAGES=C export LANG LC_MESSAGES

Actually, this isn't a remedy since bb_start starts daemons via
"su - -c" - which will reset most environment variables. If I weren't
that lazy I had replaced that su long ago (actually, I don't want
to use "su" at all, rather seteuid() within the start script).

Best regards,
Tom
-- 
----------------------------------------------------------------------------
Thomas Aeby, Kirchweg 52, 1735 Giffers, Switzerland, Tel: (+41)264180040
Internet: ae...@gr...                       PGP public key available
----------------------------------------------------------------------------

On Tue, 2006-06-13 at 21:23 +0200, Niels Baggesen wrote:
> so I MUST suggest that it is brought in line with other languages. As it
> is, it puts < > == at a lower priority than && and ||, which is quite
> surprising.

You are right. Probably this has happened accidentally, at least I 
cannot thing of one single reason why one would like to have logical
operators having priority over comparisons.

... in CVS, now ...

> Btw, your christmas feulleton included a chapter on "syntax, statements,
> operators and functions", and one on "arrays and tables". These chapters
> are missing in the developer manual at the docu site. If you have lost
> them (of course you haven't) I have them archieved ...

Thanks, indeed I still have them :-))

The developer manual is going to become a part of the main manual.
Probably I'll only add the missing chapters then.

Best regards,
Tom

P.S.: I do have noticed and noted your propositions about the plugin
page ... and also the bug causing you to use "00".
-- 
----------------------------------------------------------------------------
Thomas Aeby, Kirchweg 52, 1735 Giffers, Switzerland, Tel: (+41)264180040
Internet: ae...@gr...                       PGP public key available
----------------------------------------------------------------------------

There have been a couple of incidents on the list where df output and
others have been misinterpreted because they used local language headings
and the like.

I use the following setup:

Index: tools/bs_boot.sh
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS  /cvsroot/bigsister/bigsister/tools/bs_boot.sh,v
retrieving revision 1.2
diff -u -r1.2 bs_boot.sh
--- tools/bs_boot.sh    2 Mar 2004 15:05:44 -0000       1.2
+++ tools/bs_boot.sh    12 Jun 2006 12:16:27 -0000
@@ -17,6 +17,8 @@

 bs_instdir=3D/some-dir

+LANG=3DC LC_MESSAGES=3DC export LANG LC_MESSAGES
+
 [ -x "$bs_instdir/bin/bb_start" ] || exit 0

 exec $bs_instdir/bin/bb_start "$@"


which should prevents most problems.

/Niels

--=20
Niels Baggesen -- @home -- =C5rhus -- Denmark -- ni...@ba...
The purpose of computing is insight, not numbers  --  R W Hamming

Hi,

the operator precedence in the tester language has puzzled me lately,
so I MUST suggest that it is brought in line with other languages. As it
is, it puts < > =3D=3D at a lower priority than && and ||, which is quite
surprising.

May I suggest

Index: uxmon/Monitor/Tester.pm
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS  /cvsroot/bigsister/bigsister/uxmon/Monitor/Tester.pm,v
retrieving revision 1.40
diff -u -r1.40 Tester.pm
--- uxmon/Monitor/Tester.pm     24 Jan 2006 15:40:44 -0000      1.40
+++ uxmon/Monitor/Tester.pm     12 Jun 2006 12:16:27 -0000
@@ -940,9 +940,9 @@
     "." =3D> 3,
     "*" =3D> 2,
     "/" =3D> 2,
-    ">" =3D> 8,
-    "<" =3D> 8,
-    "=3D=3D" =3D> 8
+    ">" =3D> 4,
+    "<" =3D> 4,
+    "=3D=3D" =3D> 4
 );
 my %defaults =3D (
     "&&" =3D> 0,


Btw, your christmas feulleton included a chapter on "syntax, statements,
operators and functions", and one on "arrays and tables". These chapters
are missing in the developer manual at the docu site. If you have lost
them (of course you haven't) I have them archieved ...

/Niels

--=20
Niels Baggesen -- @home -- =C5rhus -- Denmark -- ni...@ba...
The purpose of computing is insight, not numbers  --  R W Hamming