Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-24 Thread Wes Hardaker
> "RS" == Roland Stigge <[EMAIL PROTECTED]> writes: >> /MF (I do think Debian should stop violating RFC2741) RS> Does it? Please point me to the respective areas. Taking a complete guess, because I don't know to what he was referring, does debian move the agentx socket from /var/agentx/maste

Re: MIBs in net-snmp

2006-10-20 Thread Dave Shield
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

Re: MIBs in net-snmp

2006-10-20 Thread Robert Story
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

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-18 Thread Magnus Fromreide
I am afraid that we are in violent agreement. As I understood the problem the snmp tools (not the daemons) get installed in a central location but only nonpriviliged users ever run it. In this situation it is a good thing to run any of the commands as a priviliged user, this is true as of today.

AgentX socket path (was: Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp])

2006-10-18 Thread Thomas Anders
Roland Stigge wrote: > Magnus Fromreide wrote: >> /MF (I do think Debian should stop violating RFC2741) > > Does it? Please point me to the respective areas. I think Magnus refers to the AgentX socket path. From RFC2741: 8.2.1. Well-known Values The master agent creates a well-known UNIX-dom

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-18 Thread Roland Stigge
Hi, Magnus Fromreide wrote: > The idea of a default .index is broken as you can't know what other mibs > the user might have installed, please do regenerate it as needed. User installed MIBs should be installed in a different location anyway (with the discussed system of multiple mibs dirs no pro

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Wes Hardaker
> "MF" == Magnus Fromreide <[EMAIL PROTECTED]> writes: >> How about separating index file generation code from index using >> code? This way, we could provide a snmp-update-index program to be >> called on install, package postinst, admin or user mibs updates, >> etc., and other library code

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Wes Hardaker
> "DS" == Dave Shield <[EMAIL PROTECTED]> writes: DS> Agreed. DS> At least up until the first time that a privileged account runs one of DS> the Net-SNMP tools. DS> That would then automatically generate the indexes, which would then DS> be available for later uses. Note that the default snmp

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Magnus Fromreide
On tis, 2006-10-17 at 15:19 +0200, Roland Stigge wrote: > Hi, > How about separating index file generation code from index using code? > This way, we could provide a snmp-update-index program to be called on > install, package postinst, admin or user mibs updates, etc., and other > library code wou

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Roland Stigge
Hi, Dave Shield wrote: > At least up until the first time that a privileged account runs one of > the Net-SNMP tools. > That would then automatically generate the indexes, which would then > be available for later uses. > > Actually, the one disadvantage of the patch that I supplied is that > it'

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Dave Shield
On 17/10/06, Roland Stigge <[EMAIL PROTECTED]> wrote: > In the restricted case (admin adding MIBs, daemon/program etc. running > as user with restricted permissions), it is equivalent to having no > .index files at all. ;-> Agreed. At least up until the first time that a privileged account runs on

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Roland Stigge
Hi, Dave Shield wrote: >> Yes, moving the indexes to a separate location like /var/lib/snmp is a >> reasonable compromise. A good, documented way for using multiple mibs >> locations would be nice. > > You mean something like: > >$ man snmp.conf >[snip] > mibdirs (mib-dirs|+mib-

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Dave Shield
On 17/10/06, Roland Stigge <[EMAIL PROTECTED]> wrote: > Yes, moving the indexes to a separate location like /var/lib/snmp is a > reasonable compromise. A good, documented way for using multiple mibs > locations would be nice. You mean something like: $ man snmp.conf [snip] mibdi

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Roland Stigge
Dave Shield wrote: > The index files avoid the need to do this every single time - it's an > efficiency measure. OK. > Did you have a chance to look at the patch I've put together to > address the FHS issue? Yes, moving the indexes to a separate location like /var/lib/snmp is a reasonable compro

Re: [Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Dave Shield
On 17/10/06, Roland Stigge <[EMAIL PROTECTED]> wrote: > Another idea: Is the .index file really technically necessary? It looks > like always read and regenerated (and rewritten). No - I don't believe that's true. The index file is only regenerated if it's older than the timestamp on the directory

[Fwd: Re: [Pkg-net-snmp-devel] Bug#389434: MIBs in net-snmp]

2006-10-17 Thread Roland Stigge
FYI... (btw: Thomas' email address bounced several times.) --- Begin Message --- Thomas Anders wrote: > This may be desirable for /usr/share/snmp/mibs on a fully > package-managed system, but custom builds and users with ~/.snmp/mibs > may not be prepared to do the extra work before using the MIBs,

Re: MIBs in net-snmp

2006-10-16 Thread Dave Shield
On 16/10/06, Thomas Anders <[EMAIL PROTECTED]> wrote: What's problematic (and arguably a bug) is the management of the .index file in the /usr file system. See http://sf.net/support/tracker.php?aid=1414566 http://sourceforge.net/mailarchive/message.php?msg_id=13249497 for additional details

Re: MIBs in net-snmp

2006-10-16 Thread Wes Hardaker
> "RS" == Roland Stigge <[EMAIL PROTECTED]> writes: >> What MIB files in /usr/share/snmp/mibs "can be changed"? RS> Right, maybe they need not be changed. But documentation (FAQ) describes RS> "How do I add a MIB to the tools" where the contents of the mibs dir, RS> and therefore the mibs lis

Re: MIBs in net-snmp

2006-10-16 Thread Roland Stigge
Hi Thomas, Thomas Anders wrote: > Roland Stigge wrote: >> at Debian, we noticed that the net-snmp versions 5.2 ... 5.4 don't >> properly implement the File Hierarchy Standard (FHS). Therefore, we are >> about to move /usr/share/snmp/mibs to /var/lib/snmp/mibs since it >> contains MIB files that ca

Re: MIBs in net-snmp

2006-10-16 Thread Thomas Anders
Roland Stigge wrote: > at Debian, we noticed that the net-snmp versions 5.2 ... 5.4 don't > properly implement the File Hierarchy Standard (FHS). Therefore, we are > about to move /usr/share/snmp/mibs to /var/lib/snmp/mibs since it > contains MIB files that can be changed and the .index file is eve

MIBs in net-snmp

2006-10-16 Thread Roland Stigge
forwarded 389434 [email protected] tag 389434 upstream thanks Hi Net-snmp-coders, at Debian, we noticed that the net-snmp versions 5.2 ... 5.4 don't properly implement the File Hierarchy Standard (FHS). Therefore, we are about to move /usr/share/snmp/mibs to /var/lib/snmp/mibs