5.4.pre4 is now on the download site. We're either going to a pre5 or
a rc1 from this (decision not yet made). Either way we're going into
release-candidate freeze for the main line, which means a patch needs
24 hours of review and at least positive 3 votes from knowledgeable
people on the coder
Hi,
The subagent core dumps when a regression test for restarting
master and subagent performed. In snmp_call_callbacks(), scp points is 0x10, a
non-access memory.
The following is output of gdb : (line # may not match
office net-snmp src since I added some debug lines).
I am on
> "AP" == Anders Persson <[EMAIL PROTECTED]> writes:
AP> t_uscalar_t is just a 32-bit wide type, so in this case using uint32_t
AP> should work fine. The attached patch should address the problem.
Applied.
--
Wes Hardaker
Sparta, Inc.
Bruce Shaw wrote:
> Summary:
>
> 1. We default to Perl support.
Only after a reasonable set of tests to ensure that it's likely to build/work.
If configure thinks it won't (like for most of what you mentioned here), it'll
be disabled. If it misses a significant test (I'm still not sure how many
Robert Story wrote:
> Darwin 8 & 9 don't include the freeBSD or generic system header, and thus end
> up with no INP_NEXT_SYMBOL header. This simple patch adds the symbol to the
> darwin 8 & 9 headers.
+1, as already said on IRC.
+Thomas
--
Thomas Anders (thomas.anders at blue-cable.de)
-
Darwin 8 & 9 don't include the freeBSD or generic system header, and thus end
up with no INP_NEXT_SYMBOL header. This simple patch adds the symbol to the
darwin 8 & 9 headers.
Index: include/net-snmp/system/darwin8.h
=== RCS
file: /cv
On 20/10/06, Venkata Guddeti <[EMAIL PROTECTED]> wrote:
>> How are these object implemented? Scalars, or table objects?
> These variables are implemented as scalars.
>> Which helper?
> I am sorry that I could not understand this question.
What does the code of your "init" routine look like?
On Fri, 20 Oct 2006 13:50:02 +0100 Dave wrote:
DS> On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
DS> > DS> Because 'req' starts by pointing to the first varbind (index 1).
DS> > DS> So if 'idx' is 1, then we're already at the right place.
DS> >
DS> > Ok, then shouldn't it be 'i<=idx'? Otherw
On 19/10/06, Wes Hardaker <[EMAIL PROTECTED]> wrote:
> > "TA" == Thomas Anders <[EMAIL PROTECTED]> writes:
> TA> Also, shouldn't we add a note to default_store.h/ds_agent.h
> TA> reminding people to also update the perl modules if they add
> TA> anything here?
>
> Yes, please do so. Thanks!
D
On Fri, 20 Oct 2006 20:49:27 +0100 Dave wrote:
DS> > I have coded Counter64 variables using the 'struct counter64' and filling
DS> > its members high and low. I have passed the args for
DS> > snmp_set_var_typed_value as (u_char*) and the size as 8.
DS>
DS> It's probably safer to use "sizeof(struct
On 20/10/06, Anders Persson <[EMAIL PROTECTED]> wrote:
> t_uscalar_t is just a 32-bit wide type, so in this case using uint32_t
> should work fine. The attached patch should address the problem.
+1
This communication is intended for the use of the recipient to which it is
addressed, and may cont
Dave Shield wrote:
> On 20/10/06, Anders Persson <[EMAIL PROTECTED]> wrote:
>> t_uscalar_t is just a 32-bit wide type, so in this case using uint32_t
>> should work fine. The attached patch should address the problem.
>
> +1
+1
+Thomas
--
Thomas Anders (thomas.anders at blue-cable.de)
---
On 20/10/06, Anders Persson <[EMAIL PROTECTED]> wrote:
> t_uscalar_t is just a 32-bit wide type, so in this case using uint32_t
> should work fine. The attached patch should address the problem.
+1
Dave
-
Using Tomcat but ne
On 20/10/06, Venkata Guddeti <[EMAIL PROTECTED]> wrote:
> The snmpd is actually timing out (most of the times, crash sometimes) when I
> use snmpget for two Counter64 variables.
Does querying these two variables individually work correctly?
How are these object implemented? Scalars, or table ob
BS> We have a problem within net-snmp related to computing swap resources
that
BS> makes use of a library deprecated with LARGEFILE support (which most
Perl's
BS> have compiled in).
DTS>What would be:
a) the implications
b) the feasibility
DTS>of trying to find another approach for retrievi
[ First - *please* don't mail me privately, without copying
any responses to the mailing list. I don't have the time
or inclination to offer private, unpaid, SNMP consultancy.
Keep discussions to the list, where others can both learn
and offer advice. Thanks. ]
On 20/10
There was one sentence in this message that sprung out at me:
On 20/10/06, Bruce Shaw <[EMAIL PROTECTED]> wrote:
> We have a problem within net-snmp related to computing swap resources that
> makes use of a library deprecated with LARGEFILE support (which most Perl's
> have compiled in).
What wou
t_uscalar_t is just a 32-bit wide type, so in this case using uint32_t
should work fine. The attached patch should address the problem.
Anders
Thomas Anders wrote:
I get the following build failure for CVS MAIN on Solaris 2.6:
solaris26# ./configure --enable-mfd-rewrites --with-mib-modules="
As noted in README.solaris, there have been a number of issues with Perl on
Sun over the years.
It is our policy that ./configure should work out of the box. In a variety
of cases it doesn't. This is because our current default is to
automatically include embedded Perl modules.
I would like it
2006/10/20, Thomas Anders <[EMAIL PROTECTED]>:
> Michael Tatarinov wrote:
> > 1. exclude sysfs, usbfs, binfmt_misc from hrFSTable
> > 2. set hrStorageType on loop device as hrStorageRemovableDisk in
> > hrStorageTable
> > 3. determine hrSWInstalledType for rpm in hrSWInstalledTable
>
> Thanks. Can
On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
> DS> I offer the attached patch for your delectation and delight
>
> Can you give a summary of what all this new code is supposed to do?
Moves the MIB directory index files into (a subdirectory of) the N/S persistent
directory - i.e. under
Dear all: I am Clement. I meet a program of cross compiling. when I set configure as following: ./configure --with-cc=xscale_be-gcc enable-shared=no then the following meesage shows: configure: error: cannot run C compiled program
I get the following build failure for CVS MAIN on Solaris 2.6:
solaris26# ./configure --enable-mfd-rewrites --with-mib-modules="if-mib
udp-mib tcp-mib" [...]
solaris26# make
[...]
gcc -I../../include -I/home/anders/src/net-snmp-5.4cvs/include -I.
-I../../agent -I/home/anders/src/net-snmp-5.4cvs/ag
Dear all: I am Clement. I meet a program of cross compiling. when I set configure as following: ./configure --with-cc=xscale_be-gcc enable-shared=no then the following meesage shows: configure: error: cannot run C compiled program,
If you want to cross compile, use ' --host' however , i don
On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
> DS> Because 'req' starts by pointing to the first varbind (index 1).
> DS> So if 'idx' is 1, then we're already at the right place.
>
> Ok, then shouldn't it be 'i<=idx'? Otherwise if index = 1, it will still use
> the first varbind?
But if in
On Wed, 18 Oct 2006 13:25:49 +0100 Dave wrote:
DS> On a slightly different note - should we (i.e. Net-SNMP) consider
DS> adjusting the default location of the persistent directory? A quick
DS> perusal of the FHS specs would indicate that we should perhaps be
DS> using /var/lib/net-snmp.
DS> Or i
On Fri, 20 Oct 2006 13:31:39 +0100 Dave wrote:
DS> On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
DS> > DS> +/*
DS> > DS> + * Skip to the indicated varbind
DS> > DS> + */
DS> > DS> +for ( i=2; i > DS> +req = req->next;
DS>
DS> > Why is i initialized to 2?
DS>
DS> Bec
On Thu, 19 Oct 2006 22:09:29 -0700 chris wrote:
CJ> On 19-Oct-2006, at 3:57 PM, Robert Story wrote:
CJ>
CJ> > The new hardware cpu code fails to build on my OS X box (darwin7).
CJ> >
CJ> Since I don't run Panther anymore I didn't attempt to update the
CJ> directives for that release. I recommend
On Mon, 16 Oct 2006 22:52:26 +0100 Dave wrote:
DS> On 16/10/06, Thomas Anders <[EMAIL PROTECTED]> wrote:
DS> > What's problematic (and arguably a bug) is the management of the .index
DS> > file in the /usr file system. See
DS> >
DS> > http://sf.net/support/tracker.php?aid=1414566
DS> > http://s
On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
> DS> +/*
> DS> + * Skip to the indicated varbind
> DS> + */
> DS> +for ( i=2; i DS> +req = req->next;
> Why is i initialized to 2?
Because 'req' starts by pointing to the first varbind (index 1).
So if 'idx' is 1, then
On Tue, 17 Oct 2006 11:55:07 -0700 Wes wrote:
WH> > "AS" == Avnindra Singh <[EMAIL PROTECTED]> writes:
WH>
WH> AS> Occasionally, I need values from tables implemented in subagent B in
WH> AS> subagent A.
WH>
WH> AS> Can somebody please tell a way to get such info efficiently?
Your favorite
Michael Tatarinov wrote:
> 1. exclude sysfs, usbfs, binfmt_misc from hrFSTable
> 2. set hrStorageType on loop device as hrStorageRemovableDisk in
> hrStorageTable
> 3. determine hrSWInstalledType for rpm in hrSWInstalledTable
Thanks. Can you please submit this to http://www.net-snmp.org/patches s
On Tue, 17 Oct 2006 09:48:27 +0100 Dave wrote:
DS> Index: agent/snmp_agent.c
DS> +int
DS> +netsnmp_request_set_error_idx(netsnmp_request_info *request,
DS> + int error_value, int idx)
DS> +{
...
DS> +/*
DS> + * Skip to the indicated varbind
DS> + */
DS> +
Robert Story wrote:
> On Wed, 18 Oct 2006 18:36:26 +0200 Thomas wrote:
> TA> Bruce Shaw wrote:
> TA> > When you make, it crashes with the following message:
> TA> >
> TA> > In file included from ../agent/mibgroup/mib_module_includes.h:124,
> TA> > from mib_modules.c:48:
> TA> > ./
On 20/10/06, Robert Story <[EMAIL PROTECTED]> wrote:
> I like the idea of using the same tokens, but not the same conf file. Our
> library handles application conf files, so I'd suggest they use
> audit_snmptrap.conf or some such.
I suspect that it's a deliberate decision to use the same conf file
On Wed, 18 Oct 2006 18:36:26 +0200 Thomas wrote:
TA> Bruce Shaw wrote:
TA> > When you make, it crashes with the following message:
TA> >
TA> > In file included from ../agent/mibgroup/mib_module_includes.h:124,
TA> > from mib_modules.c:48:
TA> > ./mibgroup/kernel_sunos5.h:118: warn
On Thu, 19 Oct 2006 10:16:33 -0700 Wes wrote:
WH> BS> Is this possible?
WH>
WH> I think it'd be better to attach the demon as an agentx sub agent
WH> rather than simply figure out how to parse our config files.
I like the idea of using the same tokens, but not the same conf file. Our
library hand
Hello
1. exclude sysfs, usbfs, binfmt_misc from hrFSTable
2. set hrStorageType on loop device as hrStorageRemovableDisk in hrStorageTable
3. determine hrSWInstalledType for rpm in hrSWInstalledTable
net-snmp-5.4.pre3-host-res.patch
Description: Binary data
-
On 20/10/06, Thomas Anders <[EMAIL PROTECTED]> wrote:
> > I see more stuff, but no CPU stats.
>
> What's the output of
> net-snmp-config --snmpd-module-list | tr " " "\n" | grep cpu_
>
> ? ...cpu_sysctl? The content of agent/mibgroup/hardware/cpu.h vs. the comments
> in agent/mibgroup/hardware/cpu/
39 matches
Mail list logo
