My MIB file has one main table defined (Table1) and has multiple other
tables (Table2,3,4,5,6) defined which are extensions to Table1.
So Robert . instead of me implementing sub-containers, lets say I implement
your other suggestion of adding all the fields defined in Table2,3,4,5,6 into
On fre, 2007-06-01 at 12:22 +0200, Thomas Anders wrote:
> Magnus,
>
> the configure-related changes in r16451 also broke compilation on IRIX
> 6.5. The root cause seems to be that instead of "#define
> IFNET_NEEDS_KERNEL 1" (before r16451) we now get "#undef
> NETSNMP_IFNET_NEEDS_KERNEL".
>
> I c
Hi Robert,
You are right. After 1000 traps I am not seeing any memory leaks. Is
there any document where I can get this kind of informations?
Thanks a lot for pointing out this!!!
On Fri, 2007-06-01 at 10:27 -0400, Robert Story wrote:
> On Thu, 24 May 2007 14:44:23 -0700 Navaneetham wrote:
>
Well, I went to the link you suggested and read about the Containers, however,
nothing is really making any sense to me. I am not real sure how to create a
sub-Container from an existing container.
Perhaps you could provide a simple code example of what you are talking about?
Robert Story <[E
Following is my snmpd.conf file.
+++
syslocation 192.168.0.217
rwcommunity clovis
trap2sink 127.0.0.1:2162 clovis
master yes
agentaddress 161
+++
On Fri, 2007-06-01 at 10:53 -0400, Robert Story wrote:
> On Wed, 23 May 20
On 01/06/07, Robert Story <[EMAIL PROTECTED]> wrote:
> I'm not thrilled about the idea of lots of similarly named files
> (eg 'interface.c')...
That's fair enough.
I'm not particularly thrilled about it either.
> I'd rather eliminate the sub-directory, though I'm not
> keen on that idea either
On 01/06/07, Need Help <[EMAIL PROTECTED]> wrote:
> The "--with-mib-modules" option seems to tell NetSNMP to "look for" the
> "agent/mibgroup/OC-STB-HOST-MIB.h" file only rather than create it.
That's correct.
"configure" is used for compiling the software.
"mib2c" is used for generating code.
Th
On Fri, 1 Jun 2007 08:12:56 -0700 (PDT) Need wrote:
NH> Here is the configure command I am currently using:
NH>
NH> ../configure \
NH> --prefix=/export/home/rosent1/snmp/usr/local \
NH> --srcdir=/export/home/rosent1/snmp/net-snmp-5.4 \
NH> --enable-embedded-perl \
NH> --enable-shared \
NH> --with-
On Fri, 1 Jun 2007 16:03:35 +0100 Dave wrote:
DS> The other contributory factor is that there's a lot of duplication in
DS> this naming
DS> system - particularly with the multi-level directory structure. If
DS> you've got a
DS> table SomeTable, then the files will typically be called
DS>
D
On 01/06/07, Robert Story <[EMAIL PROTECTED]> wrote:
> Another option would be to modify the subagent code to explicitly check the
> file modification time periodically, and re-read the config file via the same
> library calls the hup handling uses. then you won't need to send a signal at
> all.
U
Here is the configure command I am currently using:
../configure \
--prefix=/export/home/rosent1/snmp/usr/local \
--srcdir=/export/home/rosent1/snmp/net-snmp-5.4 \
--enable-embedded-perl \
--enable-shared \
--with-mib-modules="OC-STB-HOST-MIB"
The "--with-mib-modules" option seems to tell NetSNM
On Fri, 1 Jun 2007 07:05:59 -0700 (PDT) Need wrote:
NH> Do you recommend I generate my code using some other type of configuration
file or is it possible to implement your suggestion using the "mfd" code layout
approach?
NH>
NH> In addition I have no idea what you mean by subcontainers and filte
Hi Robert,
I have another CLI process to control enable/disable the subagent traps
configuration. Currently I made the CLI work by snmpset the trap
configuration objects and save the p tokens in the permanent subagent .conf
file. This is working successfully to turn on/off subagent traps
generatio
On 01/06/07, Robert Story <[EMAIL PROTECTED]> wrote:
> DS> Is there any reason for having (or continuing with) a separate mib2c
> DS> configuration mechanism for this, rather than using the existing
> DS>
> DS> net-snmp-config --compile-subagent
> DS>
> DS> approach?
>
> If --compile-subagent
On Fri, 1 Jun 2007 09:27:24 -0500 [EMAIL PROTECTED] wrote:
EYC> I have another CLI process to control enable/disable the subagent traps
EYC> configuration. Currently I made the CLI work by snmpset the trap
EYC> configuration objects and save the p tokens in the permanent subagent .conf
EYC> file. T
On Fri, 01 Jun 2007 16:23:18 +0200 Thomas wrote:
TA> > Did you want to trigger the re-read from withing the subagent, or
externally? I
TA> > think that the libraries will re-read all conf files on receiving a
SIGHUP.
TA>
TA> The libraries? I think the SIGHUP handling is in the applications (e.g.
On 01/06/07, Robert Story <[EMAIL PROTECTED]> wrote:
> If star is used, I believe it's only the windows folks that are affected.
Probably.
But Windows is meant to be a supported platform.
(Now who's foolish idea was that...)
> I don't suppose anyone would like my idea that 'mibgroup' be moved
On Fri, 25 May 2007 07:34:05 -0700 (PDT) Need wrote:
NH> I created the
".../agent/mibgroup/OC-STB-HOST-MIB/ocStbHostAVInterfaceTable/" directory,
issued the "mib2c" command in this directory for the
"ocStbHostAVInterfaceTable" OID node and the code was generated for me and
placed into this dire
On Fri, 1 Jun 2007 14:21:23 +0100 Dave wrote:
DS> > It should be easy to adapt this conf file for use with _any_ module
generated
DS> > by mib2c.
DS>
DS> Is there any reason for having (or continuing with) a separate mib2c
DS> configuration mechanism for this, rather than using the existing
DS>
On Wed, 23 May 2007 20:23:33 -0700 Navaneetham wrote:
N> * Problem:
N> --
N> I am using net-snmp 5.4 version. One of our test application(sub-Agent)
N> calls send_v2trap() function 100 times with minimum delay of 10 Millie-
N> Sec. snmptrapd receives more than 100 traps. Connection between
On 01/06/07, Need Help <[EMAIL PROTECTED]> wrote:
> ... my concern
> with expanding the data_context structure of Table1 with fields from Table2
> was that I thought performing a "snmpwalk" through Table1 would then display
> Table2 fields now
No.
The results returned for Table1
On Tue, 22 May 2007 20:27:34 +0100 Dave wrote:
DS> In fact, your example is simply a specific case of the more general
DS> approach. Instead of
DS>
DS> snmpwalk -W foo.x.y.z foo
DS>
DS> (with an optional starting point), this would become
DS>
DS> snmpwalk -CE {foo+1} foo.x.y.z
DS>
Being new to the whole NetSNMP structure and everything, I kind of like seeing
the "-mib" included in the directory name for the MIB.It reminds me the
directory is for MIB specific stuff rather than some generic directory.
My 2-cents of course
Robert Story <[EMAIL PROTECTED]> wrote:
On Thu, 24 May 2007 09:41:33 -0700 Wes wrote:
WH> >> I can always move them into the correct locations
WH> >> (since the individual flenames are correct), but it
WH> >> does look as if we should seriously consider shortening
WH> >> these names.
WH>
WH> TA> I'd be fine with that. "svn move" was one
Ok, I know I can change any code I want to fit my needs, however, my concern
with expanding the data_context structure of Table1 with fields from Table2 was
that I thought performing a "snmpwalk" through Table1 would then display Table2
fields now and that would be bad since Table2 fields r
On 01/06/07, Thomas Anders <[EMAIL PROTECTED]> wrote:
> The libraries? I think the SIGHUP handling is in the applications (e.g.
> apps/snmptrapd.c).
Which probably means that including the MIB module within the
main agent framework (and running as a subagent) will honour SIGHUP.
Including the modu
On Thu, 24 May 2007 14:44:23 -0700 Navaneetham wrote:
N> You are right. The bug is not explaining when that issue was happening.
N> Any idea, any one is working on this memory leak issue?
One thing to note is that the notification log mib keeps the most recent 1000
traps in memory. So there is a n
Robert Story wrote:
> On Wed, 30 May 2007 13:48:18 -0500 [EMAIL PROTECTED] wrote:
> EYC> The subagent framework is from mib2c.mfd.conf, I added
> EYC> snmpd_register_config_handler() by myself so it can read from subagent
> EYC> .conf file and save the p tokens to permanent subagent .conf file. Tha
On Thu, 31 May 2007 20:54:48 +0100 Dave wrote:
DS> FR #592109:
DS> Traditionally, the Host Resources MIB implementation has
DS> reported NFS partitions as "fixed (local) disks", although there
DS> are options to report them correctly (or skip them altogether).
DS> * Should we consider repo
On Wed, 30 May 2007 13:07:28 -0700 (PDT) Need wrote:
NH> 1) Table1 has the "ocStbHostAVInterfaceIndex" field which I believe
represents the row index in the table. Actually. I believe this index field is
used as the row index value generated within NetSNMP by doing the following:
NH>
NH>a)
Do you recommend I generate my code using some other type of configuration file
or is it possible to implement your suggestion using the "mfd" code layout
approach?
In addition I have no idea what you mean by subcontainers and filters. If you
can elaborate a bit that would be helpful. Perhap
On Wed, 30 May 2007 13:48:18 -0500 [EMAIL PROTECTED] wrote:
EYC> The subagent framework is from mib2c.mfd.conf, I added
EYC> snmpd_register_config_handler() by myself so it can read from subagent
EYC> .conf file and save the p tokens to permanent subagent .conf file. That
EYC> part is working succe
Most likely I simply interpreted some description incorrectly along the way
.
Robert Story <[EMAIL PROTECTED]> wrote: On Thu, 31 May 2007 09:39:55 +0100 Dave
wrote:
DS> On 30/05/07, Need Help wrote:
DS> > However, I do not want to give up on the MfD configuration just yet. I
was
DS> > r
On Thu, 31 May 2007 11:17:22 -0700 (PDT) Need wrote:
NH> OK, so I need to use the same index for Table2 as I used in Table1, but I
do not know how to actually do this?
Well, you've stepped out of the bounds of a 'simple' table, so Mad won't
auto-generate nice neat ready to use code. However, ther
On 01/06/07, Robert Story <[EMAIL PROTECTED]> wrote:
> DS> The MfD framework (and this framework *only*) also has the concept
> DS> of running the MIB module as a standalone agent, wrapped within a
> DS> minimal SNMP agent driver. It is unclear to me how much this
> DS> particular mode of operati
On Thu, 31 May 2007 09:39:55 +0100 Dave wrote:
DS> On 30/05/07, Need Help <[EMAIL PROTECTED]> wrote:
DS> > However, I do not want to give up on the MfD configuration just yet. I
was
DS> > reading about it and it says that it is the most complete configuration
file
DS> > offered and that existin
Thanks for the information .
Hey, I believe you might be the "Robert" who "David" was referring to who might
be able to help answer some questions regarding the "mfd" configuration. David
has helped me out quite a bit, but perhaps you might have a look at my post to
see if you can answer s
On Fri, 25 May 2007 09:33:07 +0100 Dave wrote:
DS> > I was told that the VACM might not be incorporated into the MFD
DS> > packages so it might not be able to be used easily.
DS> [...]
DS> The MfD framework (and this framework *only*) also has the concept
DS> of running the MIB module as a standalo
On Wed, 23 May 2007 11:13:37 -0700 (PDT) Need wrote:
NH> Now, I originally generated the code for each Table one at a time, but I
found it did not generate the required .h files for each new Table.
NH>
NH> For example:
NH>
NH> I went into the
".../net-snmp-5.4/agent/mibgroup/OC-STB-HOST-MIB/oc
On Fri, 18 May 2007 06:20:43 -0700 (PDT) Need wrote:
NH> Then what is the purpose of the "row_prep" procedure ? Basically, when
would i place code in the "row_prep" procedure?
The only thing the container really needs is the index values, so it can find
the right instances when a request comes i
alok bharti wrote:
> error on subcontainer '' insert (-1)
What net-snmp version are you using?
> On debugging i found that this happens for 'proxy'
> statements in my snmpd.conf file. If i remove 'proxy'
> statements, this errors disappear.
>
> morover I see memory leak while snmpd updates with
Hi,
I want my snmpd to take the update dynamically from
the snmpd.conf file.
After updating the snmpd.conf i am calling
'update_config' function to get the agent updated with
new changes in configuration.
It throws following error when updating:
error on subcontainer '' insert (-1)
error on subc
I am in the middle of writing a big discovery engine so I am mulling
over the design a bit and haven't decided whether I want to use the
sessions class...
On 6/1/07, Dave Shield <[EMAIL PROTECTED]> wrote:
> On 01/06/07, Noah Gift <[EMAIL PROTECTED]> wrote:
> > Is there an advantage to using Sessio
On 01/06/07, Noah Gift <[EMAIL PROTECTED]> wrote:
> Is there an advantage to using Session instead of just writing your
> own session? I guess most of the code I have been writing has been
> using OO python bindings and it does seem to be written in an OO
> fashion there, especially Session.
Ah -
Is there an advantage to using Session instead of just writing your
own session? I guess most of the code I have been writing has been
using OO python bindings and it does seem to be written in an OO
fashion there, especially Session.
These are some of the python examples from test:
sess =
On 01/06/07, Thomas Anders <[EMAIL PROTECTED]> wrote:
> > FR #744697:
> > * Is it worth suppressing this help output?
> > * Is it worth skipping these (irrelevant) flags in the option
> > processing?
>
> Can the presence of these options cause snmpdf et al. to malfunction?
I don't thin
On 01/06/07, Noah Gift <[EMAIL PROTECTED]> wrote:
> I would like to know more about session as well. It seems like an OO
> construct? A high level Object to inherit?
No - the Net-SNMP suite is not an object-oriented design.
A "session" is the structure used to handle the connection to a
remote
I would like to know more about session as well. It seems like an OO
construct? A high level Object to inherit?
On 6/1/07, zhangyao <[EMAIL PROTECTED]> wrote:
>
>
>
> i don't know the exact concept of session. Maybe it is useless for me. My
> work is to implement the v3
> agent.i want to know w
Dave Shield wrote:
> FR #592109:
> Traditionally, the Host Resources MIB implementation has
> reported NFS partitions as "fixed (local) disks", although there
> are options to report them correctly (or skip them altogether).
> * Should we consider reporting them as NFS mounts by default
>
Magnus,
the configure-related changes in r16451 also broke compilation on IRIX
6.5. The root cause seems to be that instead of "#define
IFNET_NEEDS_KERNEL 1" (before r16451) we now get "#undef
NETSNMP_IFNET_NEEDS_KERNEL".
I can provide more details if the problem can't be spotted easily enough.
Gupta, Aarushi wrote:
>I am trying to build Net-SNMP 5.4 on HP-UX platform .
> I tried with gcc ( 4.x & 3.x ).The issue is there is Perl module
> Which need C compiler & not gcc.
> We need to provide perl as incase we disable perl "tkmib"
> Is not working.
Perl modules (not only ours) need
Hi,
I am trying to build Net-SNMP 5.4 on HP-UX platform .
I tried with gcc ( 4.x & 3.x ).The issue is there is Perl module
Which need C compiler & not gcc.
We need to provide perl as incase we disable perl "tkmib"
Is not working.
Please suggest some means to resolve this issue
Regards,
Aar
52 matches
Mail list logo
