Xymon Mailing List Archive search

New enable/disable tool ready for testing

58 messages in this thread

list Henrik Størner · Thu, 14 Apr 2005 15:53:02 +0200 ·
I've spent the past couple of days working on the enable/disable code,
in an attempt to get rid of the maint.pl Perl script.

I have something now that I believe works well, so if you feel like
trying it out I've generated a patch for you against Hobbit 4.0.2. It
is available at http://www.hswn.dk/beta/maintenance-feature.patch

Installing is the usual:

    cd hobbit-4.0.2
    patch -p0 </tmp/maintenance-feature.patch
    make
    stop Hobbit
    make install (as root)
    start Hobbit

My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.

You can of course disable tests (a single one, some of them, or all of
them); you can re-enable tests that have been disabled; you can
schedule a disable to happen at a later time; and you can cancel
scheduled disables if they are no longer needed.

The scheduled disabling is handled internally in the Hobbit daemon, so
you no longer need "cron" running to handle it. I've never been fond
of allowing my CGI scripts to run at-jobs from a security
perspective...


Regards,
Henrik
list Lars Ebeling · Thu, 14 Apr 2005 16:07:24 +0200 (CEST) ·
Got this:
 
Not Found
The requested URL /hobbit-seccgi/hobbit-enadis.sh was not found on this server.
/Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Kevin Hanrahan · Thu, 14 Apr 2005 10:07:35 -0400 ·
I like the approach taken but can you disable ALL, ALL with a single action?
Once a month, we have a fiasco called maintenance Sunday where all the
different groups are at work taking down all their respective systems for
patching, routine maint. And whatever needs to be done. I just select All
hosts, All tests and disable for x hours. Will this still be available?

Kevin
quoted from Henrik Størner

-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] 
Sent: Thursday, April 14, 2005 9:53 AM
To: user-ae9b8668bcde@xymon.invalid
Subject: [hobbit] New enable/disable tool ready for testing


I've spent the past couple of days working on the enable/disable code, in an
attempt to get rid of the maint.pl Perl script.

I have something now that I believe works well, so if you feel like trying
it out I've generated a patch for you against Hobbit 4.0.2. It is available
at http://www.hswn.dk/beta/maintenance-feature.patch

Installing is the usual:

    cd hobbit-4.0.2
    patch -p0 </tmp/maintenance-feature.patch
    make
    stop Hobbit
    make install (as root)
    start Hobbit

My solution works somewhat differently than the old "maint.pl" script: The
enable/disable functionality is no longer a separate webpage listing all
hosts; instead, this function is moved to each hosts' "info" column page. I
did this because I have about 2000 hosts, so the list of all hosts was
practically unusable in my setup. You can see how it looks on the demo site
http://www.hswn.dk/hobbit/ - just pick one of the "info" buttons.

You can of course disable tests (a single one, some of them, or all of
them); you can re-enable tests that have been disabled; you can schedule a
disable to happen at a later time; and you can cancel scheduled disables if
they are no longer needed.

The scheduled disabling is handled internally in the Hobbit daemon, so you
no longer need "cron" running to handle it. I've never been fond of allowing
my CGI scripts to run at-jobs from a security perspective...


Regards,
Henrik


Note:  The information contained in this email and in any attachments is
intended only for the person or entity to which it is addressed and may
contain confidential and/or privileged material.  Any review,
retransmission, dissemination or other use of, or taking of any action in
reliance upon, this information by persons or entities other than the
intended recipient is prohibited.  The recipient should check this email and
any attachments for the presence of viruses.  Sender accepts no liability
for any damages caused by any virus transmitted by this email. If you have
received this email in error, please notify us immediately by replying to
the message and delete the email from your computer.  This e-mail is and any
response to it will be unencrypted and, therefore, potentially unsecure.
Thank you.  NOVA Information Systems, Inc.
list Gordon Thiesfeld · Thu, 14 Apr 2005 09:09:56 -0500 ·
quoted from Kevin Hanrahan
My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.

The demo looks good. My concern is that we often disable all tests on up to
30 servers at a time to deploy code and reboot.  Having to do each box
individually would be a real chore.
list Lars Ebeling · Thu, 14 Apr 2005 16:13:58 +0200 (CEST) ·
I have installed the new version (which) didn't work. 
But the old maint.pl is still there
 
/Lars
"Thiesfeld, Gordon" <user-708d5343dd9a@xymon.invalid> wrote:

Från: Thiesfeld, Gordon[SMTP:user-708d5343dd9a@xymon.invalid]
Skickat: den 14 april 2005 16:09:56
Till: user-ae9b8668bcde@xymon.invalid
Ämne: RE: [hobbit] New enable/disable tool ready for testing
Automatiskt vidarebefordrat av en regel
quoted from Gordon Thiesfeld

My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.

The demo looks good. My concern is that we often disable all tests on up to
30 servers at a time to deploy code and reboot. Having to do each box
individually would be a real chore.


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Lars Ebeling · Thu, 14 Apr 2005 16:15:16 +0200 (CEST) ·
Sorry for the bold and large characters , but it was just cut and paste.
 
/Lars
quoted from Lars Ebeling

lars ebeling <user-b8ea1138b488@xymon.invalid> wrote:


Från:   lars ebeling[SMTP:user-b8ea1138b488@xymon.invalid] 
Skickat:      den 14 april 2005 16:07:24 
Till:         user-ae9b8668bcde@xymon.invalid 
Ämne:     Re: [hobbit] New enable/disable tool ready for testing 
Automatiskt vidarebefordrat av en regel

Got this:
 
Not Found
The requested URL /hobbit-seccgi/hobbit-enadis.sh was not found on this server.
/Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
 

Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Paul van Eldijk · Thu, 14 Apr 2005 16:31:45 +0200 ·
quoted from Lars Ebeling
On Thursday 14 April 2005 15:53, Henrik Stoerner wrote:
I've spent the past couple of days working on the enable/disable code,
in an attempt to get rid of the maint.pl Perl script.
[snip]
My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.
The problem with this sollution is, that everybody with acces to the 
monitorpages can now enable/disable tests. We like to make information 
available to all IT-staff, but the enabling/disabling is being done by 
operations, based on RFC and calls. How to prevent others from 
enabling/disabling in this new concept?

Paul
list Tom Georgoulias · Thu, 14 Apr 2005 10:35:44 -0400 ·
quoted from Paul van Eldijk
Henrik Stoerner wrote:
My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.
Glad to see this feature!  I checked it out and have some comments about what *I'd* like to see, but this may not be what everyone else wants. I'll say it anyway.  ;)

1.  I still like having the ability to disable many hosts at once from the same page, using the pull down menu in the tool bar instead of teh info column.

2.  I found it confusing to have both the "disable now" and "schedule disable" buttons right next to each other, and the cause field over the schedule disable.

I want to log a cause for every disable/enable, scheduled or immediate.   I'd love to have a form that has separate fields for the comment and a name, so I can some what enforce my local policy of "put your name in the comment field so we know who's doing what on the systems".

Something like this:

Select enable/disable from teh tool bar, go to the

Click on the host(s) & tests you want disabled.  Since you have tons of systems and don't want to scroll through all of them, maybe you could have 2 columns, one for the group and the other for the systems in that group.

Be forced to enter both a comment and a user name to track who/what is going on.

Select from a pull down menu if you want this immediately or in the future.  if the future is selected, a new set of menus appears that allows you to schedule the start/stop times.

Push an "apply" button and everything goes into effect.

Thats my two cents.  Flame away.

Tom
list Henrik Størner · Thu, 14 Apr 2005 16:36:41 +0200 ·
quoted from Paul van Eldijk
On Thu, Apr 14, 2005 at 04:31:45PM +0200, Paul van Eldijk wrote:
My solution works somewhat differently than the old "maint.pl" script:
The enable/disable functionality is no longer a separate webpage
listing all hosts; instead, this function is moved to each hosts'
"info" column page. I did this because I have about 2000 hosts, so the
list of all hosts was practically unusable in my setup. You can see
how it looks on the demo site http://www.hswn.dk/hobbit/ - just pick
one of the "info" buttons.
The problem with this sollution is, that everybody with acces to the 
monitorpages can now enable/disable tests. We like to make information 
available to all IT-staff, but the enabling/disabling is being done by 
operations, based on RFC and calls. How to prevent others from 
enabling/disabling in this new concept?
Everyone can see what's disabled, but using the enable/disable
function still requires a separate authorization just like maint.pl
does today.

That's why you cannot disable anything on my demo page - unless you
can guess what password I use.


Regards,
Henrik
list Frédéric Mangeant · Thu, 14 Apr 2005 16:37:02 +0200 ·
quoted from Lars Ebeling
Got this:
 
Not Found

The requested URL /hobbit-seccgi/hobbit-enadis.sh was not found on this
server.
Try appliying this one _after_ maintenance-feature.patch :


--- hobbitd/Makefile    2005-04-14 15:56:34.008199160 +0200
+++ hobbitd/Makefile.new        2005-04-14 16:32:34.693529464 +0200
@@ -1,6 +1,6 @@
 PROGRAMS = hobbit.sh hobbitd hobbitd_channel hobbitd_filter
hobbitd_filestore hobbitd_history hobbitd_alert hobbitd_larrd hobbitd_sample
hobbitgraph.cgi hobbitsvc.cgi hobbit-enadis.cgi hobbit-mailack trimhistory
hobbitreports.sh
-CGISCRIPTS = hobbitgraph.sh bb-hostsvc.sh bb-histlog.sh hobbit-enadis.sh
-SECCGISCRIPTS = bb-ack.sh maint.pl
+CGISCRIPTS = hobbitgraph.sh bb-hostsvc.sh bb-histlog.sh
+SECCGISCRIPTS = bb-ack.sh maint.pl hobbit-enadis.sh
 
 LIBOBJS = ../lib/libbbgen.a


--

Frédéric Mangeant

Steria EDC Sophia Antipolis
list Henrik Størner · Thu, 14 Apr 2005 16:38:40 +0200 ·
quoted from Lars Ebeling
On Thu, Apr 14, 2005 at 04:07:24PM +0200, lars ebeling wrote:
Got this:
 
Not Found
The requested URL /hobbit-seccgi/hobbit-enadis.sh was not found on this server.
Check if this script did get installed in your secure CGI directory.
It should do that with a "make install" ...

Also, I haven't tested it if you do not use the secure/unsecure CGI
setup but have all the CGI's in one directory. That may be broken.


Henrik
list Henrik Størner · Thu, 14 Apr 2005 16:41:55 +0200 ·
quoted from Lars Ebeling
On Thu, Apr 14, 2005 at 04:13:58PM +0200, lars ebeling wrote:
I have installed the new version (which) didn't work. 
But the old maint.pl is still there
Yep - it doesn't break anything. But it might go away once I'm certain
the new stuff works.


Henrik
list Henrik Størner · Thu, 14 Apr 2005 16:44:41 +0200 ·
quoted from Frédéric Mangeant
On Thu, Apr 14, 2005 at 04:37:02PM +0200, Frédéric Mangeant wrote:
Got this:
 
Not Found

The requested URL /hobbit-seccgi/hobbit-enadis.sh was not found on this
server.
Try appliying this one _after_ maintenance-feature.patch :
Duh - of course. Thanks for spotting this, I've corrected the patch
for any new downloaders out there.


Regards,
Henrik
list Lars Ebeling · Thu, 14 Apr 2005 16:50:41 +0200 (CEST) ·
quoted from Henrik Størner

Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:


Check if this script did get installed in your secure CGI directory.
It should do that with a "make install" ...

No it wasn't. Move it there manually then everything seems to hang.
quoted from Henrik Størner

Also, I haven't tested it if you do not use the secure/unsecure CGI
setup but have all the CGI's in one directory. That may be broken.


I use cgi-secure
quoted from Lars Ebeling

Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Lars Ebeling · Thu, 14 Apr 2005 17:03:42 +0200 (CEST) ·

Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:

Try appliying this one _after_ maintenance-feature.patch :
Was it meant to be a patch here?
quoted from Henrik Størner

Duh - of course. Thanks for spotting this, I've corrected the patch
for any new downloaders out there.

I enjoy doing this, it's my hobby among other things. I'm now making dinner that shall be ready when my wife is home from work.


Regards,
quoted from Lars Ebeling
Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Henrik Størner · Sat, 16 Apr 2005 20:26:04 +0200 ·
OK, it seemed like most of you liked this new feature, but all of you
still wanted a way of disabling stuff on many hosts at once. So I've
been busy adding - essentially creating a Hobbit-version of the full
maint.pl script.

The result is an updated maintenance-feature patch, available as
always from http://www.hswn.dk/beta/maintenance-feature-v2.patch

This is against the 4.0.2 release, like the first version of the
patch.

The patch implements the same feature as the first version - you can
disable tests for a single host from the "info" page - but it adds an
extra script "hobbit-enadis.cgi" that takes over from the maint.pl
script. hobbit-enadis.cgi displays an administration page very much
like the maint.pl one; I've unashamedly stolen Tom's JavaScript code
and used it in my implementation. (I hope he doesn't mind - I'll ask
for his permission). However, to better cope with the enormous
host-list I've added the ability to filter hosts using three regular
expressions - against the hostname, the pagename of the host, and the
IP address. That makes it slightly more usable to me...


And:
quoted from Tom Georgoulias

On Thu, Apr 14, 2005 at 10:35:44AM -0400, Tom Georgoulias wrote:
2.  I found it confusing to have both the "disable now" and "schedule 
disable" buttons right next to each other, and the cause field over the 
schedule disable.
This has been fixed. I think the current display is very intuitive.
quoted from Tom Georgoulias

I want to log a cause for every disable/enable, scheduled or
 immediate. 
This is now forced via some Javascript validation, which also makes
sure you don't hit the "Apply" button without selecting hosts, tests,
entering a valid duration etc.


Regards,
Henrik
list Kevin Hanrahan · Sat, 16 Apr 2005 16:12:33 -0400 ·
OK, I installed the new patch without any problems but I get odd results
when trying to use it. Using it from the Info page:
	Microsoft IE - "A runtime error has occurred. Line 267 - Object
expected
		selecting a different server produces a different line
number but the same error

	Firefox Mozilla - clicking on the disable button does absolutely
nothing

From the Enable/Disable page:
	Seems to work fine
	The history keeps a record of who disabled and why
quoted from Kevin Hanrahan


Kevin

 
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] 
Sent: Saturday, April 16, 2005 2:26 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] New enable/disable tool ready for testing

Importance: Low
quoted from Henrik Størner

OK, it seemed like most of you liked this new feature, but all of you
still wanted a way of disabling stuff on many hosts at once. So I've
been busy adding - essentially creating a Hobbit-version of the full
maint.pl script.

The result is an updated maintenance-feature patch, available as always
from http://www.hswn.dk/beta/maintenance-feature-v2.patch

This is against the 4.0.2 release, like the first version of the patch.

The patch implements the same feature as the first version - you can
disable tests for a single host from the "info" page - but it adds an
extra script "hobbit-enadis.cgi" that takes over from the maint.pl
script. hobbit-enadis.cgi displays an administration page very much like
the maint.pl one; I've unashamedly stolen Tom's JavaScript code and used
it in my implementation. (I hope he doesn't mind - I'll ask for his
permission). However, to better cope with the enormous host-list I've
added the ability to filter hosts using three regular expressions -
against the hostname, the pagename of the host, and the IP address. That
makes it slightly more usable to me...


And:

On Thu, Apr 14, 2005 at 10:35:44AM -0400, Tom Georgoulias wrote:
2.  I found it confusing to have both the "disable now" and "schedule 
disable" buttons right next to each other, and the cause field over 
the schedule disable.
This has been fixed. I think the current display is very intuitive.

I want to log a cause for every disable/enable, scheduled or  
immediate.
This is now forced via some Javascript validation, which also makes sure
you don't hit the "Apply" button without selecting hosts, tests,
entering a valid duration etc.


Regards,
Henrik
list Henrik Størner · Sat, 16 Apr 2005 22:34:24 +0200 ·
quoted from Kevin Hanrahan
On Sat, Apr 16, 2005 at 04:12:33PM -0400, kevin hanrahan wrote:
OK, I installed the new patch without any problems but I get odd results
when trying to use it. Using it from the Info page:
	Microsoft IE - "A runtime error has occurred. Line 267 - Object
expected
		selecting a different server produces a different line
number but the same error
Could you try doing a "View source" and let me know what's on that
line and the surrounding 3-5 lines ?

I've tried it with IE before releasing and it worked for me,
obviously. I've validated the HTML, but this is probably something
with the Javascript code.

	Firefox Mozilla - clicking on the disable button does absolutely
nothing
Probably the same issue. I saw something similar during testing.


Just to rule out any ghosts - did you do a "make clean" before
building Hobbit after the patch was applied ? 


Henrik
list Michael Dunne · Sat, 16 Apr 2005 16:40:55 -0400 ·
Greetings,

When I attempt to apply the patch I receive the following error:

gcc -g -O -D_REENTRANT -DBIND_8_COMPAT=1 -DDarwin -DMAXMSG=32768
-DBBDPORTNUMBER=1984 -I. -I/Users/hobbit/Desktop/hobbit-4.0.2/include
-DHOBBITD=1 -I. -I../include    -c -o headfoot.o headfoot.c
headfoot.c:25:18: pcre.h: No such file or directory
headfoot.c:53: error: parse error before '*' token
headfoot.c:53: warning: data definition has no type or storage class
headfoot.c:54: error: parse error before '*' token
headfoot.c:54: warning: data definition has no type or storage class
headfoot.c:55: error: parse error before '*' token
headfoot.c:55: warning: data definition has no type or storage class
headfoot.c: In function `sethostenv_filter':
headfoot.c:118: error: `PCRE_CASELESS' undeclared (first use in this
function)
headfoot.c:118: error: (Each undeclared identifier is reported only once
headfoot.c:118: error: for each function it appears in.)
headfoot.c:118: warning: assignment makes pointer from integer without a
cast
headfoot.c:119: warning: assignment makes pointer from integer without a
cast
headfoot.c:120: warning: assignment makes pointer from integer without a
cast
make[1]: *** [headfoot.o] Error 1
make: *** [lib-build] Error 2


I did a "make clean" before building. This is on an OS X system running
10.3.8 server. 


Any suggestions?


Thanks in advance.

Mike
quoted from Henrik Størner

On 4/16/05 4:34 PM, "Henrik Stoerner" <user-ce4a2c883f75@xymon.invalid> wrote:
On Sat, Apr 16, 2005 at 04:12:33PM -0400, kevin hanrahan wrote:
OK, I installed the new patch without any problems but I get odd results
when trying to use it. Using it from the Info page:
Microsoft IE - "A runtime error has occurred. Line 267 - Object
expected
selecting a different server produces a different line
number but the same error
Could you try doing a "View source" and let me know what's on that
line and the surrounding 3-5 lines ?

I've tried it with IE before releasing and it worked for me,
obviously. I've validated the HTML, but this is probably something
with the Javascript code.

Firefox Mozilla - clicking on the disable button does absolutely
nothing
Probably the same issue. I saw something similar during testing.


Just to rule out any ghosts - did you do a "make clean" before
building Hobbit after the patch was applied ?


Henrik

list Kevin Hanrahan · Sat, 16 Apr 2005 17:05:53 -0400 ·
OK,
 Did a
	 make clean
	make
	make install

Tried again and got the same result from IE. Here is the surrounding
code:


 </SELECT>
              </td></tr>
            </table> 
         </td>
(267)      </tr>
      <tr> <td align=center> <input name=apply type="button"
onClick="validateDisable(this.form)" value="Apply"></td> </tr>
   </table>
</td>
</table>
<input name="hostname" type=hidden value="zeus">


Also, Firefox still does nothing.
quoted from Michael Dunne


Kevin


-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] 
Sent: Saturday, April 16, 2005 4:34 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] New enable/disable tool ready for testing
Importance: Low

On Sat, Apr 16, 2005 at 04:12:33PM -0400, kevin hanrahan wrote:
OK, I installed the new patch without any problems but I get odd 
results when trying to use it. Using it from the Info page:
	Microsoft IE - "A runtime error has occurred. Line 267 - Object 
expected
		selecting a different server produces a different line
number but 
the same error
Could you try doing a "View source" and let me know what's on that line
and the surrounding 3-5 lines ?

I've tried it with IE before releasing and it worked for me, obviously.
I've validated the HTML, but this is probably something with the
Javascript code.

	Firefox Mozilla - clicking on the disable button does absolutely
nothing
Probably the same issue. I saw something similar during testing.


Just to rule out any ghosts - did you do a "make clean" before building
Hobbit after the patch was applied ? 


Henrik
list Henrik Størner · Sat, 16 Apr 2005 23:16:53 +0200 ·
quoted from Michael Dunne
On Sat, Apr 16, 2005 at 04:40:55PM -0400, Michael Dunne wrote:
When I attempt to apply the patch I receive the following error:

gcc -g -O -D_REENTRANT -DBIND_8_COMPAT=1 -DDarwin -DMAXMSG=32768
-DBBDPORTNUMBER=1984 -I. -I/Users/hobbit/Desktop/hobbit-4.0.2/include
-DHOBBITD=1 -I. -I../include    -c -o headfoot.o headfoot.c
headfoot.c:25:18: pcre.h: No such file or directory
Ah - right, the PCRE include-directory is not being passed to this
compile-command.

Apply this patch (on top of the one you installed), and I think it
should compile OK.


Regards,
Henrik
-------------- next part --------------
--- build/Makefile.rules	2005/03/24 07:22:32	1.165
+++ build/Makefile.rules	2005/04/16 21:14:13
@@ -56,7 +56,7 @@
 	CC="$(CC)" CFLAGS="$(CFLAGS)" RPATHOPT="$(RPATHOPT)" SSLLIBS="$(SSLLIBS)" NETLIBS="$(NETLIBS)" BBHOME="$(BBHOME)" $(MAKE) -C build all
 
 lib-build:
-	CC="$(CC)" CFLAGS="$(CFLAGS)" RPATHOPT="$(RPATHOPT)" SSLFLAGS="$(SSLFLAGS)" SSLINCDIR="$(SSLINCDIR)" SSLLIBS="$(SSLLIBS)" NETLIBS="$(NETLIBS)" BBTOPDIR="$(BBTOPDIR)" BBLOGDIR="$(BBLOGDIR)" BBHOSTNAME="$(BBHOSTNAME)" BBHOSTIP="$(BBHOSTIP)" BBHOSTOS="$(BBHOSTOS)" $(MAKE) -C lib all
+	CC="$(CC)" CFLAGS="$(CFLAGS)" RPATHOPT="$(RPATHOPT)" PCREINCDIR="$(PCREINCDIR)" SSLFLAGS="$(SSLFLAGS)" SSLINCDIR="$(SSLINCDIR)" SSLLIBS="$(SSLLIBS)" NETLIBS="$(NETLIBS)" BBTOPDIR="$(BBTOPDIR)" BBLOGDIR="$(BBLOGDIR)" BBHOSTNAME="$(BBHOSTNAME)" BBHOSTIP="$(BBHOSTIP)" BBHOSTOS="$(BBHOSTOS)" $(MAKE) -C lib all
 
 common-build: lib-build
 	CC="$(CC)" CFLAGS="$(CFLAGS)" RPATHOPT="$(RPATHOPT)" SSLFLAGS="$(SSLFLAGS)" SSLINCDIR="$(SSLINCDIR)" SSLLIBS="$(SSLLIBS)" NETLIBS="$(NETLIBS)" BBHOME="$(BBHOME)" $(MAKE) -C common all
--- lib/Makefile	2005/04/03 15:31:58	1.17
+++ lib/Makefile	2005/04/16 21:13:33
@@ -17,6 +17,9 @@
 digest.o: digest.c
 	$(CC) $(CFLAGS) $(SSLFLAGS) $(SSLINCDIR) -c -o $@ digest.c
 
+headfoot.o: headfoot.c
+	$(CC) $(CFLAGS) $(PCREINCDIR) -c -o $@ headfoot.c
• environ.o: environ.c
 	$(CC) $(CFLAGS) -DBBTOPDIR=\"$(BBTOPDIR)\" -DBBLOGDIR=\"$(BBLOGDIR)\" -DBBHOSTNAME=\"$(BBHOSTNAME)\" -DBBHOSTIP=\"$(BBHOSTIP)\" -DBBHOSTOS=\"$(BBHOSTOS)\" -c -o $@ environ.c
list Henrik Størner · Sat, 16 Apr 2005 23:26:27 +0200 ·
quoted from Kevin Hanrahan
On Sat, Apr 16, 2005 at 05:05:53PM -0400, kevin hanrahan wrote:
(267)      </tr>
      <tr> <td align=center> <input name=apply type="button"
onClick="validateDisable(this.form)" value="Apply"></td> </tr>
If you could send me the full HTML document, I'd be interested to see
if it is equally broken when I view it. This smells of a browser issue
- which IE version are you using ?


Just as an experiment, could you try changing line 348 of
hobbitd/hobbitsvc-info.c ? It currently is

    addtobuffer(buf, buflen, "      <tr> <td align=center> <input name=apply type=\"button\" onClick=\"validateDisable(this.form)\" value=\"Apply\"></td> </tr>\n");

Try changing it to

    addtobuffer(buf, buflen, "      <tr> <td align=center> <input name=apply type=\"submit\" value=\"Apply\"></td> </tr>\n");

i.e. change the type from "button" to "submit", and remove the
"onClick" stuff.

If that makes it work, then it's the Javascript form validation that
is broken - maint.pl has a config item to disable that, so maybe I
should just mimic this.


Regards,
Henrik
list Kevin Hanrahan · Sat, 16 Apr 2005 18:49:18 -0400 ·
Henrik,
 That change did the job. It is now working with both IE and Firefox 
quoted from Henrik Størner

Kevin

-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] 
Sent: Saturday, April 16, 2005 5:26 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] New enable/disable tool ready for testing
Importance: Low

On Sat, Apr 16, 2005 at 05:05:53PM -0400, kevin hanrahan wrote:
(267)      </tr>
      <tr> <td align=center> <input name=apply type="button"
onClick="validateDisable(this.form)" value="Apply"></td> </tr>
If you could send me the full HTML document, I'd be interested to see if
it is equally broken when I view it. This smells of a browser issue
- which IE version are you using ?


Just as an experiment, could you try changing line 348 of
hobbitd/hobbitsvc-info.c ? It currently is

    addtobuffer(buf, buflen, "      <tr> <td align=center> <input
name=apply type=\"button\" onClick=\"validateDisable(this.form)\"
value=\"Apply\"></td> </tr>\n");

Try changing it to

    addtobuffer(buf, buflen, "      <tr> <td align=center> <input
name=apply type=\"submit\" value=\"Apply\"></td> </tr>\n");

i.e. change the type from "button" to "submit", and remove the "onClick"
stuff.

If that makes it work, then it's the Javascript form validation that is
broken - maint.pl has a config item to disable that, so maybe I should
just mimic this.


Regards,
Henrik
list Lars Ebeling · Sun, 17 Apr 2005 11:46:07 +0200 (CEST) ·
If I start from scratch i.e. from hobbit-4.0.2.tar.gz. Which patches and in which order shall I apply.
quoted from Lars Ebeling

 
Regards

Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Henrik Størner · Sun, 17 Apr 2005 16:46:24 +0200 ·
quoted from Lars Ebeling
On Sun, Apr 17, 2005 at 11:46:07AM +0200, lars ebeling wrote:
If I start from scratch i.e. from hobbit-4.0.2.tar.gz. Which patches
 and in which order shall I apply.
1) hobbit-4.0.2-setup.patch
2) hobbit-4.0.2-acknowledge.patch
3) maintenance-feature-v2.patch

The first two are available at http://www.hswn.dk/hobbitsw/, the last
one at http://www.hswn.dk/beta/


Regards,
Henrik
list Henrik Størner · Sun, 17 Apr 2005 21:46:53 +0200 ·
quoted from Kevin Hanrahan
On Sat, Apr 16, 2005 at 06:49:18PM -0400, kevin hanrahan wrote:
 Henrik Stoerner wrote:
  Just as an experiment, could you try changing line 348 of
  hobbitd/hobbitsvc-info.c ? It currently is
 That change did the job. It is now working with both IE and Firefox 
I'm puzzled why you see this problem - both IE and Firefox handle the
Javascript stuff with no problems here. So I'm still interested to a)
hear what browser versions you're using, and b) get the full info-page
HTML that triggers this bug.

Anyway, I've added a couple of options to the hobbitsvc.cgi tool, so
you can disable the javascript validation, or completely remove the
"disable" form from the info-column page.


Regards,
Henrik
list Kevin Hanrahan · Sun, 17 Apr 2005 23:20:46 -0400 ·
Henrik,

Here are the versions of the browsers. I also tested it with two other
computers in my house with the same results. Has anyone else reported
problems?...or has anyone else tried the new feature?


 Version: 6.0.2800.1106.xpsp2.040919-1003
Cipher Strength: 128-bit 
Product ID:55274-640-0110937-23745
Update Versions:; SP1; Q330994; Q824145; Q832894; Q837009; Q831167;
Q823353; Q867801; 


Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041107
Firefox/1.0
   

Regards
quoted from Henrik Størner

KEvin

  
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] 
Sent: Sunday, April 17, 2005 3:47 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] New enable/disable tool ready for testing
Importance: Low

On Sat, Apr 16, 2005 at 06:49:18PM -0400, kevin hanrahan wrote:
 Henrik Stoerner wrote:
  Just as an experiment, could you try changing line 348 of
  hobbitd/hobbitsvc-info.c ? It currently is
 That change did the job. It is now working with both IE and Firefox
I'm puzzled why you see this problem - both IE and Firefox handle the
Javascript stuff with no problems here. So I'm still interested to a)
hear what browser versions you're using, and b) get the full info-page
HTML that triggers this bug.

Anyway, I've added a couple of options to the hobbitsvc.cgi tool, so you
can disable the javascript validation, or completely remove the
"disable" form from the info-column page.


Regards,
Henrik
list Oliver Bassett · Mon, 18 Apr 2005 17:04:23 +1200 ·
Kevin,

Might be a silly question but you don't have JavaScript disabled on those
browsers do you?

Regards
Oliver Bassett
quoted from Kevin Hanrahan
-----Original Message-----
From: kevin hanrahan [mailto:user-7211a77fa045@xymon.invalid]
Sent: Monday, 18 April 2005 3:21 p.m.
To: 'user-ae9b8668bcde@xymon.invalid'
Subject: RE: [hobbit] New enable/disable tool ready for testing


Henrik,

Here are the versions of the browsers. I also tested it with two other
computers in my house with the same results. Has anyone else reported
problems?...or has anyone else tried the new feature?


 Version: 6.0.2800.1106.xpsp2.040919-1003
Cipher Strength: 128-bit Product ID:55274-640-0110937-23745
Update Versions:; SP1; Q330994; Q824145; Q832894; Q837009; Q831167;
Q823353; Q867801; 

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.5) Gecko/20041107
Firefox/1.0
   

Regards

KEvin

  -----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] Sent: Sunday, April 17, 2005 3:47 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] New enable/disable tool ready for testing
Importance: Low

On Sat, Apr 16, 2005 at 06:49:18PM -0400, kevin hanrahan wrote:
 Henrik Stoerner wrote:
  Just as an experiment, could you try changing line 348 of
  hobbitd/hobbitsvc-info.c ? It currently is
 That change did the job. It is now working with both IE and Firefox
I'm puzzled why you see this problem - both IE and Firefox handle the
Javascript stuff with no problems here. So I'm still interested to a)
hear what browser versions you're using, and b) get the full info-page
HTML that triggers this bug.

Anyway, I've added a couple of options to the hobbitsvc.cgi tool, so you
can disable the javascript validation, or completely remove the
"disable" form from the info-column page.


Regards,
Henrik


##############################################################
################################################

This e-mail message has been scanned for Viruses and Content and cleared by NetIQ MailMarshal at Gen-i Limited ##############################################################
################################################
#############################################################################
This communication, including any attachments, is confidential.
If you are not the intended recipient, you should not read it - please contact me immediately, destroy it, and do not copy or use any part of this communication or disclose anything about it.
Thank You.

Please note that this communication does not designate an information system
for the purposes of the NZ Electronic Transactions Act 2002.

This email was scanned and cleared by NetIQ MailMarshal at Gen-i Limited.
#############################################################################
list Lars Ebeling · Mon, 18 Apr 2005 07:29:31 +0200 (CEST) ·
quoted from Henrik Størner

Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:
 

On Sun, Apr 17, 2005 at 11:46:07AM +0200, lars ebeling wrote:
If I start from scratch i.e. from hobbit-4.0.2.tar.gz. Which patches
and in which order shall I apply.
1) hobbit-4.0.2-setup.patch
2) hobbit-4.0.2-acknowledge.patch
3) maintenance-feature-v2.patch

The first two are available at http://www.hswn.dk/hobbitsw/, the last
one at http://www.hswn.dk/beta/

Now I have done this and still everything hangs when I try to disable from info page.
quoted from Lars Ebeling

 
Regards

Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Tom Georgoulias · Mon, 18 Apr 2005 10:51:59 -0400 ·
Henrik Stoerner wrote:

First of all, I like the new enadis tool.  Thanks for putting out the new, non-perl version so soon.
quoted from Kevin Hanrahan
The patch implements the same feature as the first version - you can
disable tests for a single host from the "info" page - but it adds an
extra script "hobbit-enadis.cgi" that takes over from the maint.pl
script. 
This may be a local problem, but after I built hobbit with this new patch in place my old maint.pl remained and the only way I could access the new script was by manually typing in the correct path URL.
quoted from Kevin Hanrahan

hobbit-enadis.cgi displays an administration page very much
like the maint.pl one; I've unashamedly stolen Tom's JavaScript code
and used it in my implementation. (I hope he doesn't mind - I'll ask
for his permission). However, to better cope with the enormous
host-list I've added the ability to filter hosts using three regular
expressions - against the hostname, the pagename of the host, and the
IP address. That makes it slightly more usable to me...
I like the new enadis page layout much more than the last.  However, I do have one comment--there is no easy way to enable several hosts at once, they still have to be done one by one.  Can this be changed to allow group enables?

Tom
list Henrik Størner · Mon, 18 Apr 2005 17:00:43 +0200 ·
quoted from Tom Georgoulias
On Mon, Apr 18, 2005 at 10:51:59AM -0400, Tom Georgoulias wrote:
The patch implements the same feature as the first version - you can
disable tests for a single host from the "info" page - but it adds an
extra script "hobbit-enadis.cgi" that takes over from the maint.pl
script. 
This may be a local problem, but after I built hobbit with this new patch in place my old maint.pl remained and the only way I could access the new script was by manually typing in the correct path URL.
"make install" doesn't update the menu-file. Change the
~/server/www/menu/menu_items.js file to link to hobbit-enadis.sh
instead of maint.pl
I like the new enadis page layout much more than the last.  However, I do have one comment--there is no easy way to enable several hosts at once, they still have to be done one by one.  Can this be changed to allow group enables?
The host selection listbox lets you select multiple hosts. Just hold
down <ctrl> while you click on the hosts you want to disable.


Regards,
Henrik
list Tom Georgoulias · Mon, 18 Apr 2005 11:22:10 -0400 ·
quoted from Henrik Størner
Henrik Stoerner wrote:
The host selection listbox lets you select multiple hosts. Just hold
down <ctrl> while you click on the hosts you want to disable.
Oh, I was referring to enabling them after they've already been disabled.  The ones in the "Currently disabled tests" block.  They have "enable" buttons next to each host, but no obvious way to do more than one at a time.

Tom
list Paul D. Backer · Mon, 18 Apr 2005 12:19:46 -0500 ·
BB would include the URL of the test that caused the alert in the BBALPHAMSG environment variable.  Could this be added to hobbit.

FYI - the BBCOLORLEVEL environment variable isn't listed in the on-line documentation (http://www.hswn.dk/hobbit/help/hobbit-alerts.html#scripts).

Thanks,

Paul
list Henrik Størner · Mon, 18 Apr 2005 23:26:15 +0200 ·
quoted from Paul D. Backer
On Mon, Apr 18, 2005 at 12:19:46PM -0500, Paul D. Backer wrote:
BB would include the URL of the test that caused the alert in the
 BBALPHAMSG environment variable.  Could this be added to hobbit.

FYI - the BBCOLORLEVEL environment variable isn't listed in the
 on-line documentation
 (http://www.hswn.dk/hobbit/help/hobbit-alerts.html#scripts).
Thanks, the attached patch should fix both issues. Save it to a file,
then

   cd hobbit-4.0.2
   patch -p0 </tmp/hobbit-4.0.2-alertscript.patch
   make

Shutdown hobbit, run "make install" as root, and restart hobbit.


Regards,
Henrik
-------------- next part --------------
--- hobbitd/do_alert.c	2005/03/27 06:58:46	1.60
+++ hobbitd/do_alert.c	2005/04/18 21:20:37
@@ -1203,6 +1203,10 @@
 		addtobuffer(&buf, &buflen, info);
 		addtobuffer(&buf, &buflen, msg_data(alert->pagemessage));
 		addtobuffer(&buf, &buflen, "\n");
+		sprintf(info, "See %s%s/bb-hostsvc.sh?HOSTSVC=%s.%s\n", 
+			xgetenv("BBWEBHOST"), xgetenv("CGIBINURL"), 
+			commafy(alert->hostname->name), alert->testname->name);
+		addtobuffer(&buf, &buflen, info);
 		MEMUNDEFINE(info);
 		return buf;
 
--- docs/hobbit-alerts.html	2005/02/15 12:57:33	1.6
+++ docs/hobbit-alerts.html	2005/04/18 21:22:28
@@ -146,6 +146,7 @@
 BBALPHSMSG variable for the script.</p>
 <p>The full list of environment variables provided to scripts are as follows:</p>
 <table width="80%" align="center" summary="Paging script environment variables">
+	<tr><th align="left" valign="top">BBCOLORLEVEL</th><td>The current color of the status</td></tr>
 	<tr><th align="left" valign="top">BBALPHAMSG</th><td>The full text of the status log triggering the alert</td></tr>
 	<tr><th align="left" valign="top">ACKCODE</th><td>The "cookie" that can be used to acknowledge the alert</td></tr>
 	<tr><th align="left" valign="top">RCPT</th><td>The recipient, from the SCRIPT entry</td></tr>
list Lars Ebeling · Tue, 19 Apr 2005 09:33:07 +0200 (CEST) ·
Hi all
 
I stopped and started Apache for an other reason. Then I saw that a test was disabled. The disabling for the test had been hanging for hours, and I had forgotten it. 
 
So stopping and starting Apache fix the disabling, but that I can't do everytime I disable a test from info page.

Best regards
Lars

lars ebeling <user-b8ea1138b488@xymon.invalid> wrote:
Datum: Mon, 18 Apr 2005 07:29:31 +0200 (CEST)
Från: lars ebeling <user-b8ea1138b488@xymon.invalid>
Till: hobbit <user-ae9b8668bcde@xymon.invalid>
Ämne: Fwd: Re: [hobbit] New enable/disable tool ready for testing


Från:   lars ebeling[SMTP:user-b8ea1138b488@xymon.invalid] 
Skickat:      den 18 april 2005 07:29:31 
Till:         hobbit 
Ämne:     Fwd: Re: [hobbit] New enable/disable tool ready for testing 
quoted from Lars Ebeling
Automatiskt vidarebefordrat av en regel


Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote: 
 

On Sun, Apr 17, 2005 at 11:46:07AM +0200, lars ebeling wrote:
If I start from scratch i.e. from hobbit-4.0.2.tar.gz. Which patches
and in which order shall I apply.
1) hobbit-4.0.2-setup.patch
2) hobbit-4.0.2-acknowledge.patch
3) maintenance-feature-v2.patch

The first two are available at http://www.hswn.dk/hobbitsw/, the last
one at http://www.hswn.dk/beta/

Now I have done this and still everything hangs when I try to disable from info page.

 
Regards

Lars


Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
 

Hobbithobbyist

"If you run UNIX and you don't have a UPS, you should see a psychiatrist...."
--Byte Magazine (years ago)
list Paul D. Backer · Tue, 19 Apr 2005 16:00:07 -0500 ·
I have hobbit 4.0.2 compiled and running on Solaris 10 with the following patches:

hobbit-4.0.2-setup.patch
hobbit-4.0.2-acknowledge.patch

If I try and apply this patch:

hobbit-4.0.2-maintenance-feature-v2.patch

I then get the follow compile error:

gcc -g -O2 -Wall -Wno-unused -D_REENTRANT -DHAVE_RPCENT -DSunOS -DMAXMSG=32768 -DBBDPORTNUMBER=1984 -I. -I/opt/hobbit/hobbit-4.0.2/include -DHOBBITD=1 -c -o bb-webpage.o bb-webpage.c
gcc -g -O2 -Wall -Wno-unused -D_REENTRANT -DHAVE_RPCENT -DSunOS -DMAXMSG=32768 -DBBDPORTNUMBER=1984 -I. -I/opt/hobbit/hobbit-4.0.2/include -DHOBBITD=1 -o bb-webpage bb-webpage.o ../lib/libbbgen.a -L/opt/sfw/lib -lpcre
Undefined                       first referenced
 symbol                             in file
recv                                ../lib/libbbgen.a(sendmsg.o)
getservbyname                       ../lib/libbbgen.a(sendmsg.o)
gethostbyname                       ../lib/libbbgen.a(sendmsg.o)
socket                              ../lib/libbbgen.a(sendmsg.o)
getsockopt                          ../lib/libbbgen.a(sendmsg.o)
connect                             ../lib/libbbgen.a(sendmsg.o)
inet_aton                           ../lib/libbbgen.a(sendmsg.o)
inet_ntoa                           ../lib/libbbgen.a(sendmsg.o)
shutdown                            ../lib/libbbgen.a(sendmsg.o)
ld: fatal: Symbol referencing errors. No output written to bb-webpage
collect2: ld returned 1 exit status
gmake[1]: *** [bb-webpage] Error 1
gmake[1]: Leaving directory `/opt/hobbit/hobbit-4.0.2/bbdisplay'
gmake: *** [bbdisplay-build] Error 2


I use /opt/sfw/bin/patch to apply patches with no error.  I use /opt/sfw/bin/gmake and /opt/sfw/bin/gcc for the compile.

I had maintenance-feature.patch installed and compiled at on time but started from scratch to install the maintenance-feature-v2.patch.


Thanks,

Paul
list Henrik Størner · Tue, 19 Apr 2005 23:42:34 +0200 ·
quoted from Paul D. Backer
On Tue, Apr 19, 2005 at 04:00:07PM -0500, Paul D. Backer wrote:
gcc -g -O2 -Wall -Wno-unused -D_REENTRANT -DHAVE_RPCENT -DSunOS -DMAXMSG=32768 -DBBDPORTNUMBER=1984 -I. -I/opt/hobbit/hobbit-4.0.2/include -DHOBBITD=1 -o bb-webpage bb-webpage.o ../lib/libbbgen.a -L/opt/sfw/lib -lpcre
Undefined                       first referenced
 symbol                             in file
recv                                ../lib/libbbgen.a(sendmsg.o)
There's a missing "$(NETLIBS)" in the bbdisplay/Makefile entry for
bb-webpage.

Around line 67 or so of bbdisplay/Makefile you'll find this:

bb-webpage: $(WEBPAGEOBJS) ../lib/libbbgen.a
        $(CC) $(CFLAGS) -o $@ $(WEBPAGEOBJS) ../lib/libbbgen.a $(PCRELIBS)

Just add $(NETLIBS) after the $(PCRELIBS), like it is on the other
lines around there.


Regards,
Henrik
list Tom Georgoulias · Wed, 20 Apr 2005 21:55:59 -0400 ·
quoted from Kevin Hanrahan
Henrik Stoerner wrote:
OK, it seemed like most of you liked this new feature, but all of you
still wanted a way of disabling stuff on many hosts at once. So I've
been busy adding - essentially creating a Hobbit-version of the full
maint.pl script.
I have a question about the new hobbit-native enable/disable tool, but I haven't had enough time to test it out for myself to know the answer.

One of the downsides to the way the current enable/disable maint.pl works is the way the trend charts do not continue using data collected from clients after the tests are disabled.  For example if I disable disk tests for a server, the BB client still sends the data to hobbit and the latest values are displayed, but those values aren't put into the graphs.  As a result, I have large blank gaps in spots when a test was disabled.  I can understand the presence of these gaps when a client is offline, but if I'm getting the data but just don't want be bothered with pager alerts from a server issue I'm already aware of, shouldn't that be OK?

Tom
list Henrik Størner · Thu, 21 Apr 2005 09:45:32 +0200 ·
quoted from Tom Georgoulias
On Wed, Apr 20, 2005 at 09:55:59PM -0400, Tom Georgoulias wrote:
One of the downsides to the way the current enable/disable maint.pl works is the way the trend charts do not continue using data collected from clients after the tests are disabled.  For example if I disable disk tests for a server, the BB client still sends the data to hobbit and the latest values are displayed, but those values aren't put into the graphs.  As a result, I have large blank gaps in spots when a test was disabled.  I can understand the presence of these gaps when a client is offline, but if I'm getting the data but just don't want be bothered with pager alerts from a server issue I'm already aware of, shouldn't that be OK?
It should, and I actually thought Hobbit was already working that
way. But no - I forgot to send "blue" status reports through to the
RRD update routine.

Patch is attached.

BTW, I expect to release a 4.0.3 version over the week-end with the
patches that have accumulated, and the new enable/disable tool. Before
Lars objects - I think I have found out what why it is hanging for
you, so I'll send you a patch later today for testing.


Regards,
Henrik
-- 
Henrik Storner
-------------- next part --------------
--- hobbitd/hobbitd_larrd.c	2005/03/25 21:13:41	1.17
+++ hobbitd/hobbitd_larrd.c	2005/04/21 07:40:58
@@ -12,7 +12,7 @@
 /*                                                                            */
 /*----------------------------------------------------------------------------*/
 -static char rcsid[] = "$Id: hobbitd_larrd.c,v 1.17 2005/03/25 21:13:41 henrik Exp $";
+static char rcsid[] = "$Id: hobbitd_larrd.c,v 1.18 2005/04/21 07:40:46 henrik Exp $";
  #include <stdio.h>
 #include <string.h>
@@ -175,6 +175,7 @@
 			  case COL_GREEN:
 			  case COL_YELLOW:
 			  case COL_RED:
+			  case COL_BLUE: /* Blue is OK, because it only arrives here when an update is sent */
 				tstamp = atoi(metadata[1]);
 				sender = metadata[2];
 				hostname = metadata[4];
list Frédéric Mangeant · Thu, 21 Apr 2005 11:29:58 +0200 ·
Hi Henrik
BTW, I expect to release a 4.0.3 version over the week-end with the patches that have accumulated, and the new enable/disable tool. Before Lars objects - I think I have found out what why it is hanging for you, so I'll send you a patch later today for testing.
Do you also plan to release some day bbgen 3.6 (bbgen 3.5 with the 5 patches
available here : http://www.hswn.dk/hobbitsw/bbgen/) ?

Thanks for your answer.

Regards,

-- 

Frédéric Mangeant
list Tom Georgoulias · Thu, 21 Apr 2005 10:55:56 -0400 ·
quoted from Henrik Størner
Henrik Stoerner wrote:
It should, and I actually thought Hobbit was already working that
way. But no - I forgot to send "blue" status reports through to the
RRD update routine.

Patch is attached.
Thanks for the patch.  I put it, and all the others from the list (including the new enadis tool) for 4.0.2 into production today.  So far everything works as expected.  Glad to be free of the maint.pl shortcomings.
quoted from Frédéric Mangeant
BTW, I expect to release a 4.0.3 version over the week-end with the
patches that have accumulated, and the new enable/disable tool.
Will the 4.0.3 enadis tool have any new features that aren't already present in maintenance-feature-v2.patch?

My server says hobbit 4.0.3, so I wonder if I'll even need to upgrade to the official 4.0.3 release....

Tom
list Henrik Størner · Thu, 21 Apr 2005 17:52:46 +0200 ·
quoted from Tom Georgoulias
On Thu, Apr 21, 2005 at 10:55:56AM -0400, Tom Georgoulias wrote:
BTW, I expect to release a 4.0.3 version over the week-end with the
patches that have accumulated, and the new enable/disable tool.
Will the 4.0.3 enadis tool have any new features that aren't already 
present in maintenance-feature-v2.patch?
There are some differences. The hobbit-enadis code that picks up the
data when you post an enable/disable request is buggy - it works for
most people, but not all.
quoted from Tom Georgoulias

My server says hobbit 4.0.3, so I wonder if I'll even need to upgrade to 
the official 4.0.3 release....
Oops - must have missed putting in the "4.0.2m2" version number
sometime. I would suggest upgrading, just to be sure that there are no
already-solved bugs lurking in your installation.


Regards,
Henrik
list Paul D. Backer · Thu, 21 Apr 2005 11:27:49 -0500 ·

This small enhancement will put an icon on the location bar of your browser as well as on the tab, if you browser supports tabs (such as Firefox).

First, start by inserting the following HTML tag inside the <HEAD> ... </HEAD> section of each header file in the web directory:

<link rel="shortcut icon" href="&BBSKIN/ico/favicon-&BBBACKGROUND.ico">

Second, and last, unzip the attached file into the www/gifs directory.  You should have a ico directory in the gifs directory containing seven files.

That it.  The location of the icon files maybe could be somewhere more suitable but this was the easiest for me.  The HTML tag maybe doesn't need to be added to every header file, I'm too busy to figure it out.  I used the recent.gifs for the icons, if you want others this website, http://www.html-kit.com/favicon/ can convert almost any image into an icon compatible with this process.

Paul
Attachments (1)
list Tom Georgoulias · Fri, 22 Apr 2005 08:29:15 -0400 ·
quoted from Henrik Størner
Henrik Stoerner wrote:
There are some differences. The hobbit-enadis code that picks up the
data when you post an enable/disable request is buggy - it works for
most people, but not all.
Something else we've noticed after using the new enadis tool is just how big the webpage gets once you disable all of the tests on a handful of systems.  Each test for each hosts gets its own block, and if you have 5 or so systems offline, each with all 11 tests disabled, well....the page is just huge.  Enabling those systems currently disabled is also a bit more cumbersome than before, since you have to click the enable button for each system.test.  Of course you can get around that by going to the info column for a system that you want enabled and select the ALL button, but an equivalent does not exist for the enadis.sh page.

Just my 2 cents.

Tom
list Henrik Størner · Fri, 22 Apr 2005 14:52:30 +0200 ·
quoted from Tom Georgoulias
On Fri, Apr 22, 2005 at 08:29:15AM -0400, Tom Georgoulias wrote:
Henrik Stoerner wrote:
There are some differences. The hobbit-enadis code that picks up the
data when you post an enable/disable request is buggy - it works for
most people, but not all.
Something else we've noticed after using the new enadis tool is just how big the webpage gets once you disable all of the tests on a handful of systems.  Each test for each hosts gets its own block, and if you have 5 or so systems offline, each with all 11 tests disabled, well....the page is just huge.
I thought about removing the details of each disabled test from the
page itself, and just put them in a pop-up window that appears when
you let the mouse "hover" above the host.test text. 
My current version also by default applies a "page" filter to the
enadis page, so it includes only hosts that are present on the page
you invoked it from (you can of course clear the filter and get all
hosts).
 Enabling those systems currently disabled is also a bit more cumbersome than before, since you have to click the enable button for each system.test.  Of course you can get around that by going to the info column for a system that you want enabled and select the ALL button, but an equivalent does not exist for the enadis.sh page.
Should it be
- a button to enable all tests for a single host ?
- a button to enable all tests for all hosts ?
- a checkbox so you can check off those host+test combos you want to
  enable ?
- all of the above ?


Henrik
list Frédéric Mangeant · Fri, 22 Apr 2005 15:22:47 +0200 ·
Hi all
quoted from Henrik Størner
Enabling those systems currently disabled is also a bit more 
cumbersome than before, since you have to click the enable 
button for each system.test.  Of course you can get around
that by going to the info column for a system that you want
enabled and select the ALL button, but an equivalent does not
exist for the enadis.sh page.
Should it be
- a button to enable all tests for a single host ?
- a button to enable all tests for all hosts ?
- a checkbox so you can check off those host+test combos you want to
  enable ?
- all of the above ?
I'd like this behavior :
- a button to enable all tests for a single host using hobbit-enadis.sh, as
with the "old" maint.pl
- a checkbox to choose any host+test combo using the "info" column, as it is
now.

I like the idea of using hobbit-enadis.sh for a "global" view, and the
"info" column for a more specific one.

-- 

Frédéric Mangeant
list Asif Iqbal · Fri, 22 Apr 2005 11:09:57 -0400 ·
quoted from Frédéric Mangeant
On Fri, Apr 22, 2005 at 02:52:30PM, Henrik Stoerner wrote:
On Fri, Apr 22, 2005 at 08:29:15AM -0400, Tom Georgoulias wrote:
Henrik Stoerner wrote:
 Enabling those systems currently disabled is also a bit 
more cumbersome than before, since you have to click the enable button 
for each system.test.  Of course you can get around that by going to the 
info column for a system that you want enabled and select the ALL 
button, but an equivalent does not exist for the enadis.sh page.
Should it be
- a button to enable all tests for a single host ?
- a button to enable all tests for all hosts ?
- a checkbox so you can check off those host+test combos you want to
  enable ?
- all of the above ?
All of the above. 

Also is it possible to send one email per submit instead of per
host.service?

In other words if I disable ALL or some services for ALL/Some/One
host(s), it would be nice to receive one email per action/submit instead
of receiving one email for each host.service.

Old maint.pl used to work like that, means one email only per action and
not per host.service

Just a suggestion and not critical

Thanks

-- 
Asif Iqbal
PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu
"..there are two kinds of people: those who work and those who take the credit...try
 to be in the first group;...less competition there."  - Indira Gandhi
list Tom Georgoulias · Fri, 22 Apr 2005 14:06:35 -0400 ·
quoted from Asif Iqbal
Henrik Stoerner wrote:
Should it be
- a button to enable all tests for a single host ?
- a button to enable all tests for all hosts ?
- a checkbox so you can check off those host+test combos you want to
  enable ?
- all of the above ?
Been thinking about this and at the risk of overcomplicating teh interface, I think that all of the above are needed.

My users are more likely to do all their mangement of enabling/disabling hosts from the pull down menu, not the info column.  Having full management options available on that page is key to making things simple to remember.

Tom
list Henrik Størner · Sun, 24 Apr 2005 15:50:40 +0000 (UTC) ·
quoted from Paul D. Backer
In <1114100868.14206.59.camel at laie> "Paul D. Backer" <user-0e53b970b042@xymon.invalid> writes:
This small enhancement will put an icon on the location bar of your
browser as well as on the tab, if you browser supports tabs (such as
Firefox).
I like this :-) Added for the next release.


Thanks,
Henrik
list David Gore · Fri, 29 Apr 2005 02:40:58 +0000 ·
Henrik,

Do you have plans to add back the username like in the maint.pl?  I 
would like to see who is disabling tests.

conn

Disabled by: unknown @ some.ip.addr.ess
Reason: WE can't do anything to resolve this issue

Until: Fri Apr 29 13:53:45 2005

~David
list Henrik Størner · Fri, 29 Apr 2005 07:57:37 +0200 ·
quoted from David Gore
On Fri, Apr 29, 2005 at 02:40:58AM +0000, David Gore wrote:
Do you have plans to add back the username like in the maint.pl?  I 
would like to see who is disabling tests.

conn

Disabled by: unknown @ some.ip.addr.ess
Reason: WE can't do anything to resolve this issue
The default setup puts the enable/disable on a webpage that is
password-protected, and it then automatically picks up the username 
you used to login.

If you really want to know who is disabling stuff, perhaps that was
worth considering ?


Henrik
list David Gore · Fri, 29 Apr 2005 13:17:29 +0000 ·
quoted from Henrik Størner
Henrik Stoerner wrote:
On Fri, Apr 29, 2005 at 02:40:58AM +0000, David Gore wrote:
Do you have plans to add back the username like in the maint.pl?  I would like to see who is disabling tests.

conn

Disabled by: unknown @ some.ip.addr.ess
Reason: WE can't do anything to resolve this issue

The default setup puts the enable/disable on a webpage that is
password-protected, and it then automatically picks up the username you used to login.

If you really want to know who is disabling stuff, perhaps that was
worth considering ?
I suppose I do not know what you mean.  I didn't disable anything.  I have added a password via Apache for the entire Hobbit site when our security people complained.  I cannot say I have see a password request for this specific page.

I will add security for the cgi-secure directory as well, if that is what you mean.  It many evoke some complaining about two passwords for hobbit we will see.


~David
Henrik

list Allan Marillier · Fri, 29 Apr 2005 09:47:39 -0400 ·
This may be a stupid oversight, but I am confused. I installed hobbit-4.0.2-1 yesterday on a server that was previously running BB. I took down the BB processes, completed the install and brought across my bb-hosts file. For now, I didn't bring across any of the history. Everything runs fine (better actually!), except that I don't get any updates for CPU, disk. memory, messages and processes. The clients are running just as they were before. When I look at the other server which is still running BB, I see the updates coming across the way they should.

The hobbit server was previously configured as a failover BB server, which I never managed to get truly failing over and assuming the BBPAGER role anyway. In my bb-hosts I have both servers defined as BBNET and BBDISPLAY, but only one defined as BBPAGER.

I'd appreciate any pointers on what I'm overlooking.
Thanks.
list David Gore · Fri, 29 Apr 2005 14:14:27 +0000 ·
quoted from David Gore
David Gore wrote:
Henrik Stoerner wrote:
On Fri, Apr 29, 2005 at 02:40:58AM +0000, David Gore wrote:
Do you have plans to add back the username like in the maint.pl?  I would like to see who is disabling tests.

conn

Disabled by: unknown @ some.ip.addr.ess
Reason: WE can't do anything to resolve this issue

The default setup puts the enable/disable on a webpage that is
password-protected, and it then automatically picks up the username you used to login.

If you really want to know who is disabling stuff, perhaps that was
worth considering ?

I suppose I do not know what you mean.  I didn't disable anything.  I have added a password via Apache for the entire Hobbit site when our security people complained.  I cannot say I have see a password request for this specific page.

I will add security for the cgi-secure directory as well, if that is what you mean.  It many evoke some complaining about two passwords for hobbit we will see.


~David
I suppose I do not know how to fix this.  The enable/disable doesn't pick up the login for the Hobbit site and I cannot get apache to ask for a second password on the enable/disable page.  Perhaps I have the apache config wrong?


Alias /hobbit  "/export/home/hobbit/server/www"
<Directory "/export/home/hobbit/server/www/">
     Options Indexes FollowSymLinks Includes MultiViews
     Order allow,deny
     Allow from all
     AuthUserFile /export/home/hobbit/server/etc/hobbitpasswd
     AuthType Basic
     AuthName "Hobbit Network and Application Monitoring"
     Require valid-user
</Directory>

ScriptAlias /hobbit-cgi "/www/hobbit/cgi-bin"
<Directory "/www/cgi-bin/hobbit">
     AllowOverride None
     Options ExecCGI Includes
     Order allow,deny
     Allow from all
</Directory>

ScriptAlias /hobbit-seccgi/ "/www/hobbit/cgi-secure/"
<Directory "/www/cgi-secure">
     AllowOverride None
     Options ExecCGI Includes
     Order allow,deny
     Allow from all
     AuthUserFile /export/home/hobbit/server/etc/hobbitpasswd
     AuthType Basic
     AuthName "Hobbit Administration"
     Require user dwgore
</Directory>


~David
list Henrik Størner · Fri, 29 Apr 2005 17:14:17 +0200 ·
quoted from Allan Marillier
On Fri, Apr 29, 2005 at 09:47:39AM -0400, user-e3a6ebbee6cd@xymon.invalid wrote:
This may be a stupid oversight, but I am confused. I installed hobbit-4.0.2-1 yesterday on a server that was previously running BB. I took down the BB processes, completed the install and brought across my bb-hosts file. For now, I didn't bring across any of the history. Everything runs fine (better actually!), except that I don't get any updates for CPU, disk. memory, messages and processes. The clients are running just as they were before. When I look at the other server which is still running BB, I see the updates coming across the way they should.
Check the bb-hosts file used by the BB clients - they have their own
file, with a definition of which server is the BBDISPLAY server.
The hobbit server was previously configured as a failover BB server, which I never managed to get truly failing over and assuming the BBPAGER role anyway. In my bb-hosts I have both servers defined as BBNET and BBDISPLAY, but only one defined as BBPAGER.
I think your client bb-hosts file has the old BB server listed as
BBDISPLAY, instead of the Hobbit server.

You can have both, e.g. if you have the client installed in
/usr/local/bb/bbc1.9e-btf/ (the client $BBHOME setting) then in the
etc/bb-hosts file you can have

     10.0.0.1   bbserver.foo.com     # BBDISPLAY BBPAGER
     10.0.0.2   hobbitserver.foo.com # BBDISPLAY BBPAGER

The clients will then send their status updates to both servers.

Note: You need to restart the client after changing the bb-hosts file.


Henrik
list Allan Marillier · Fri, 29 Apr 2005 12:17:17 -0400 ·
I checked again just to be sure, but I had already set all clients over a year ago to have both servers listed as BBDISPLAY and BBNET, but only one as BBPAGER. I thought that was all I needed. The hostname and IP address have not changed - all I did was bring BB down on one of the two servers, install hobbit and copy across the bb-hosts and bring hobbit up.

Everything else worked like a dream, except my ssh test, which under BB I had to put in as ssh:s. Under hobbit that kept flipping between yellow (something like an unexpected ssh connect ting I think) and green . Once I removed :s from the ssh test that was also perfect.


user-ce4a2c883f75@xymon.invalid (Henrik Stoerner) 04/29/2005 11:14 AM
Please respond to
user-ae9b8668bcde@xymon.invalid


To
user-ae9b8668bcde@xymon.invalid
cc

Subject
Re: [hobbit] Disk / CPU / memory monitoring
quoted from Henrik Størner


On Fri, Apr 29, 2005 at 09:47:39AM -0400, user-e3a6ebbee6cd@xymon.invalid wrote:
This may be a stupid oversight, but I am confused. I installed hobbit-4.0.2-1 yesterday on a server that was previously running BB. I took down the BB processes, completed the install and brought across my bb-hosts file. For now, I didn't bring across any of the history. Everything runs fine (better actually!), except that I don't get any updates for CPU, disk. memory, messages and processes. The clients are running just as they were before. When I look at the other server which 
is 
still running BB, I see the updates coming across the way they should.
Check the bb-hosts file used by the BB clients - they have their own
file, with a definition of which server is the BBDISPLAY server.
The hobbit server was previously configured as a failover BB server, 
which 
I never managed to get truly failing over and assuming the BBPAGER role anyway. In my bb-hosts I have both servers defined as BBNET and 
BBDISPLAY, 
but only one defined as BBPAGER.
I think your client bb-hosts file has the old BB server listed as
BBDISPLAY, instead of the Hobbit server.

You can have both, e.g. if you have the client installed in
/usr/local/bb/bbc1.9e-btf/ (the client $BBHOME setting) then in the
etc/bb-hosts file you can have

     10.0.0.1   bbserver.foo.com     # BBDISPLAY BBPAGER
     10.0.0.2   hobbitserver.foo.com # BBDISPLAY BBPAGER

The clients will then send their status updates to both servers.

Note: You need to restart the client after changing the bb-hosts file.


Henrik
list Henrik Størner · Sun, 1 May 2005 22:10:44 +0200 ·
quoted from David Gore
On Fri, Apr 29, 2005 at 01:17:29PM +0000, David Gore wrote:
Henrik Stoerner wrote:
On Fri, Apr 29, 2005 at 02:40:58AM +0000, David Gore wrote:
Do you have plans to add back the username like in the maint.pl?  I >>would like to see who is disabling tests.

conn

Disabled by: unknown @ some.ip.addr.ess
Reason: WE can't do anything to resolve this issue
The default setup puts the enable/disable on a webpage that is
password-protected, and it then automatically picks up the username >you used to login.
I suppose I do not know what you mean.  I didn't disable anything.  I have added a password via Apache for the entire Hobbit site when our security people complained.  I cannot say I have see a password request for this specific page.

I will add security for the cgi-secure directory as well, if that is what you mean.  It many evoke some complaining about two passwords for hobbit we will see.
As long as you use the same authentication setup for the cgi-secure
directory as for the rest of the Hobbit site, your users should not see
a second password request - the browser sends it automatically.


Henrik
list Henrik Størner · Sun, 1 May 2005 22:16:51 +0200 ·
quoted from David Gore
On Fri, Apr 29, 2005 at 02:14:27PM +0000, David Gore wrote:
I suppose I do not know how to fix this.  The enable/disable doesn't pick up the login for the Hobbit site and I cannot get apache to ask for a second password on the enable/disable page.  Perhaps I have the apache config wrong?


Alias /hobbit  "/export/home/hobbit/server/www"
<Directory "/export/home/hobbit/server/www/">
    Options Indexes FollowSymLinks Includes MultiViews
    Order allow,deny
    Allow from all
    AuthUserFile /export/home/hobbit/server/etc/hobbitpasswd
    AuthType Basic
    AuthName "Hobbit Network and Application Monitoring"
    Require valid-user
</Directory>

ScriptAlias /hobbit-cgi "/www/hobbit/cgi-bin"
<Directory "/www/cgi-bin/hobbit">
Is this Directory name correct ? By default, Hobbit puts the cgi
directory in the same top-level directory as the "server" directory, so I would expect that to be

<Directory "/export/home/hobbit/cgi-bin/hobbit">
    AllowOverride None
    Options ExecCGI Includes
    Order allow,deny
    Allow from all
</Directory>

and of course the same change for the cgi-secure directory.


Henrik