> On Tue, 18 Oct 2005 13:52:52 -0400, Robert Story <[EMAIL PROTECTED]> said:
Robert> Question to core developers: do we have a policy on byte
Robert> order? If not, we should, as this is the second time
Robert> endian-ness has bit us recently.
All the data types should be returned in local by
> On Mon, 17 Oct 2005 14:39:15 +0200, Radek Vokál <[EMAIL PROTECTED]> said:
Radek> man snmptrapd.conf mentions that the traphandle syntax is
Radek> traphandle OID|default PROGRAM [ARGS...]
Radek> But the example traphandle line in the man page mentions neither an OID
or the word 'default'. I
> On 19 Oct 2005 09:46:46 +0530, Krishna Mohan <[EMAIL PROTECTED]> said:
Krishna> traphandle default /usr/local/bin/traptoemail/ [EMAIL PROTECTED]
^
^
> On Tue, 18 Oct 2005 13:06:23 -0400, Robert Story <[EMAIL PROTECTED]> said:
Robert> On Tue, 18 Oct 2005 15:28:17 +0100 Dave wrote:
DS> I'd suggest that the more important thing here is avoiding an
DS> agent crash. We can live with a minor memory leak in the short
DS> term. (We've done so pl
> On Tue, 18 Oct 2005 13:17:52 -0400, Robert Story <[EMAIL PROTECTED]> said:
TA> It can't do any
TA> harm on other platforms than Solaris (since it lives in kernel_sunos5.c)
TA> and looks like a highly important fix for Solaris users monitoring their
TA> interfaces. We've had several compla
Dear Wes Hardaker
Thanq for ur response.
I've added traptoemail in my snmptrapd.conf file like this :
traphandle default /usr/local/bin/traptoemail/ [EMAIL PROTECTED]
Here I prefer defaults :- -f FROM root -s SMTPSERVER localhost, b'cos i'm executing snmpd & snmptrapd on the same host.
Yes. I'm testing Solaris 8 right now.
>What solaris versions have you tested with?
>Bruce, can you check the patch on the myriad of versions you have,
especially
>the older ones, and see if everything still works?
This communication is intended for the use of the recipient to which it is
addre
On Tue, 18 Oct 2005 12:06:06 -0700 Wes wrote:
WH>
WH> Robert> Without this patch, a subagent which has successfully
WH> Robert> connected to the master agent will encounter a double-free
WH> Robert> during shutdown, and will most likely crash. Not as critical
WH> Robert> as a crash during normal o
On Tue, 18 Oct 2005 21:31:02 +0200 Niels wrote:
NB> Correct. The function has all the time been documented as returning zero
NB> for ok. At some point in time somebody has used it with incorrect checks
NB> for ok returns. I have been over all uses and the complete code to make
NB> sure all agrees o
On Tue, 18 Oct 2005 11:07:14 -0700 Srini wrote:
SK> > On Mon, 17 Oct 2005 11:56:17 -0700 Srini wrote:
SK> > SK> From the mib impletation code we return the values in
SK> > host order and
SK> > SK> we believe snmp stack should format the values in network
SK> > byte order
SK> > SK> and send on w
>Correct. The function has all the time been documented as returning zero
>for ok. At some point in time somebody has used it with incorrect checks
>for ok returns. I have been over all uses and the complete code to make
>sure all agrees on the head documentation. To make things even more
>complica
On Tue, Oct 18, 2005 at 01:17:52PM -0400, Robert Story wrote:
> I was inclined to agree until I took a look at the code in question. We've
> gone from setting an error for (rc < 0), to (rc == 0) and now (rc != 0). This
> looks like either people don't know what they are doing, or the error
> condit
> On Tue, 18 Oct 2005 13:28:00 -0400, Robert Story <[EMAIL PROTECTED]> said:
Robert> Without this patch, a subagent which has successfully
Robert> connected to the master agent will encounter a double-free
Robert> during shutdown, and will most likely crash. Not as critical
Robert> as a crash
comments inline
- srini
> -Original Message-
> From: Robert Story [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, October 18, 2005 10:53 AM
> To: Srini Kode (skode)
> Cc: [email protected]; Dave Shield; Wes
> Subject: Re: Endian - ness issue
>
> On Mon, 17 Oct 2005 11:56
Robert Story wrote:
We've long had double free's in the shutdown code, and one of the causes was
trying to clear up the client args in the callback structures. There is code
that I added in all 3 branches that loops through all the callbacks and clears
a given client arg from any callback that is
On Mon, 17 Oct 2005 11:56:17 -0700 Srini wrote:
SK> From the mib impletation code we return the values in host order and we
SK> believe snmp stack should format the values in network byte order and
SK> send on wire.
What are the types of the object involved here?
Question to core developers: do
On 17 Oct 2005 22:05:07 +0530 Kandasamy wrote:
K>I have to send sanity notification to multiple managers who are
K> registered for trap/inform. This sanity notification needs to be send to
K> each manager at different time interval.
K>
K> Let say if two managers A and B registered for sanity
We've long had double free's in the shutdown code, and one of the causes was
trying to clear up the client args in the callback structures. There is code
that I added in all 3 branches that loops through all the callbacks and clears
a given client arg from any callback that is using it. This is cal
On Tue, 18 Oct 2005 18:54:34 +0200 Thomas wrote:
TA> It can't do any
TA> harm on other platforms than Solaris (since it lives in kernel_sunos5.c)
TA> and looks like a highly important fix for Solaris users monitoring their
TA> interfaces. We've had several complaints in the past on the lists.
>It can't do any harm on other platforms than Solaris (since it lives in
kernel_sunos5.c)
>and looks like a highly important fix for Solaris users monitoring their
>interfaces. We've had several complaints in the past on the lists.
>(Anyone still listening? We could benefit from your testing!)
On Tue, 18 Oct 2005 15:28:17 +0100 Dave wrote:
DS> I'd suggest that the more important thing here is avoiding an
DS> agent crash. We can live with a minor memory leak in the short
DS> term. (We've done so plenty of times before!)
I agree... That's 4/4 so far, so go for it..
--
Robert Story; NE
Niels Baggesen wrote:
The problem is that getKstatInt isn't consistent in selecting return
values, and some places expects nonzero to be failure (as documented),
and others expect zero to be failure. And the function isn't even
consistent internally.
The problems stems from a patch that Wes appl
On Tue, 2005-10-18 at 06:58 -0700, Wes Hardaker wrote:
> Dave> The proposed patch looks reasonably straightforward, so ought to
> Dave> be safe enough. I'm not sure whether it introduces a memory leak
> Dave> (should the PDU be freed or not?), but that's less of a problem than
> Dave> an agent cra
> On Tue, 18 Oct 2005 10:31:58 +0100, Dave Shield <[EMAIL PROTECTED]> said:
Dave> The proposed patch looks reasonably straightforward, so ought to
Dave> be safe enough. I'm not sure whether it introduces a memory leak
Dave> (should the PDU be freed or not?), but that's less of a problem than
> On 18 Oct 2005 19:08:23 +0530, Krishna Mohan <[EMAIL PROTECTED]> said:
Krishna> "traphandle default /usr/bin/krsna.pl"
Krishna> It's a perl script which gives message "u have recved a trap"
Krishna> infinite times.
Can you tell us more about what it's doing? It certainly should
work. See
Dear All,
I'm able to receive traps using snmptrapd in /var/log/messages. However, I'd like snmptrapd to invoke my external program so as to alert me on receipt of trap.
In snmptrapd.conf i have written this directive
"traphandle default /usr/bin/krsna.pl"
It's a perl script which gives me
On Mon, Oct 17, 2005 at 11:19:15PM +0200, Thomas Anders wrote:
> Niels Baggesen wrote:
> >Update of /cvsroot/net-snmp/net-snmp/agent/mibgroup
> >In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv22582/agent/mibgroup
> >
> >Modified Files:
> > Tag: V5-2-patches
> > kernel_sunos5.c
> >L
On Tue, 2005-10-18 at 13:25 +0100, jillala srinivas wrote:
> Iam developing an application that should send Traps/notifications to
> LoriotPro snmp manager.
You mean like 'snmptrap' ?
> I Just want to write an application by using the net-snmp tools,
> How can I?
See 'apps/snmptrap.c'
Dave
Iam developing an application that should send Traps/notifications to LoriotPro snmp manager. I Just want to write an application by using the net-snmp tools, How can I? plz send the necessary information
thanx and regds
srinivas
Yahoo! India Matrimony: Find your partner now.
On Tue, 2005-10-18 at 08:30 -0300, Jordan Janeiro Lopes da Silva wrote:
> why the agent calls the get_{first,next} of the result table if the
> command snmptable was explicity for the control table?
Remember that there's no such low-level SNMP request as "snmptable".
Fundamentally, the only reque
Robert,
I found the problem of the error. When I called snmptable with the control
table empty, the agent calls get_first and get_next of the control table and
the get_{first,next} of the result table. In the result table, I missed a line
that didn´t test NULL, so "segmentation fault". Now I´ve got
On Mon, 2005-10-17 at 21:34 -0700, Wes Hardaker wrote:
> The following bug report documents a case where AgentX sub-agents
> sending a trap that hit a inform receiver that is down crash the main
> net-snmp agent:
I'd call that as a show-stopper.
Perhaps not under the strict definitions that I pro
On Tue, 2005-10-18 at 10:44 +0200, Thomas Anders wrote:
> Do we agree that documentation/example updates do not
> fall under these rules and can go in immediately?
Standalone documentation updates can go in right up
to the last minute. Documentation embedded within
code files should be OK up unt
Wes Hardaker wrote:
The following patch fixes this but certainly is not a change that
should be made lightly. I'm very interested in other opinions about
it, especially in light of the fact it should go into 5.2.2:
I have yet to think about possible drawbacks of this particular code,
but as f
Wes Hardaker wrote:
To reiterate, a change (even small) going into a cvs tree after the first
RC candidate has been released should:
1) have a post to -coders asking for acceptability, describing the
why's and what not.
2) show the patch, assuming a reasonable posting size
3) asking for suppo
35 matches
Mail list logo
