From owner-agentx@dorothy.bmc.com Wed Mar 1 09:00:14 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA01608 for ; Wed, 1 Mar 2000 09:00:12 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id HAA05173; Wed, 1 Mar 2000 07:57:22 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id FAA09756 for agentx-list; Wed, 1 Mar 2000 05:52:21 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id FAA09751 for ; Wed, 1 Mar 2000 05:52:18 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id HAA04011 for ; Wed, 1 Mar 2000 07:52:53 -0600 (CST) Received: by lespaul.process.com with Internet Mail Service (5.5.2650.21) id ; Wed, 1 Mar 2000 08:52:47 -0500 Message-ID: <63D30D6E10CFD11190A90000F805FE8602455A94@lespaul.process.com> From: Richard Whalen To: "'Alexandre Briani Kieling'" Cc: "'agentx@dorothy.peer.com'" Subject: RE: agentx X pure snmp Date: Wed, 1 Mar 2000 08:52:47 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-1" Sender: owner-agentx@dorothy.bmc.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list I recently added Agent X support to our "traditional" monolithic agent. We chose to do this because it would make it easier for others to add SNMP support in the areas that they need it. With Agent X the support for these additional MIBs is dynamic - SNMP does not need to be restarted. Also, a company that decides to put SNMP monitoring in their product can do it in a standard way, rather than having to work with many different SNMP agent interfaces. Also, if the product is a software product, then there would need to be communication between the SNMP agent and the actual product. Agent X puts the management closer to the what is being managed, where there is a better chance of having all of the information be available and current. As Agent X gets more popular I hope that a "standard" set of client routines becomes available that will reduce the amount of work necessary to exchange Agent X PDUs. -----Original Message----- From: Alexandre Briani Kieling [mailto:briani@led.ufsc.br] Sent: Tuesday, February 29, 2000 7:39 PM To: agentx@dorothy.peer.com Cc: ucd-snmp@ucd-snmp.ucdavis.edu Subject: agentx X pure snmp I have been studying the agentX protocol for six months and I am starting to write a document about this technology. But I can not just talk about theory. I need to validate what I say. I mean, if I say that extensible agents are better than monolitic agents, I have to prove this. Is the implementation of extensible agents easier than the traditional way of implement snmp agents? Is there a way to determine the time spent developing subagents and snmp agents. Is it true that many subagents have better performance than one big snmp agent? Is it better to develop a new module and compile it with a existent snmp agent like ucd-snmp or is it better to develop a extensible agent using agentX? I am not sure about it. Is there any way I can prove that agentX is better? What I am asking here is a suggestion about how I can compare the two technologies and validate the results. Could anyone who developed a extensible agent tell me what type of information the agent was managing? Is there any situation where only extensible agents can do the job? I do not mean to bother you with so many questions. I would really appreciate any hint. Thanks for your time. -- Alexandre Briani Kieling Laboratory of Distance Learning Federal University of Santa Catarina Brazil From owner-agentx@dorothy.bmc.com Wed Mar 1 10:00:43 2000 Received: from tangelo.bmc.com (fw-us-hou-1.bmc.com [198.207.223.250]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA03655 for ; Wed, 1 Mar 2000 10:00:40 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id IAA23053; Wed, 1 Mar 2000 08:55:12 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id GAA09850 for agentx-list; Wed, 1 Mar 2000 06:52:09 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id GAA09845 for ; Wed, 1 Mar 2000 06:52:05 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id IAA21982 for ; Wed, 1 Mar 2000 08:52:39 -0600 (CST) Received: from bnatale (ppp9.acec.com [38.249.211.62]) by relay1.acec.com (8.9.3/8.9.3) with ESMTP id JAA21185; Wed, 1 Mar 2000 09:51:45 -0500 (EST) Message-Id: <4.2.2.20000301095324.00b3e9d8@plymouth.acec.com> X-Sender: bnatale@plymouth.acec.com X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.2 Date: Wed, 01 Mar 2000 09:59:12 -0500 To: Richard Whalen From: Bob Natale Subject: RE: agentx X pure snmp Cc: agentx@dorothy.peer.com In-Reply-To: <63D30D6E10CFD11190A90000F805FE8602455A94@lespaul.process.c om> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-agentx@dorothy.bmc.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list At 3/1/2000:08:52 AM, Richard Whalen wrote: Hi Richard, >I recently added Agent X support to our "traditional" monolithic agent. We >chose to do this because it would make it easier for others to add SNMP >support in the areas that they need it. With Agent X the support for these >additional MIBs is dynamic - SNMP does not need to be restarted. Also, a >company that decides to put SNMP monitoring in their product can do it in a >standard way, rather than having to work with many different SNMP agent >interfaces. > >Also, if the product is a software product, then there would need to be >communication between the SNMP agent and the actual product. Agent X puts >the management closer to the what is being managed, where there is a better >chance of having all of the information be available and current. Excellent summary of the primary rationale behind the extensible agent approach...with AgentX adding the standardization factor that makes the approach more economically appealing to all parties. >As Agent X gets more popular I hope that a "standard" set of client routines >becomes available that will reduce the amount of work necessary to exchange >Agent X PDUs. Producing such an API (or equivalent approach(es)) is an optional work item of the AgentX WG. Now that the second rev of the protocol and the formalizatioin of the initial MIB are complete (in RFCs 2741 and 2742, respectively), perhaps the group should consider whether we want to tackle the API issue or not. I welcome input on that question. Moreover, if you could draft a brief outline of your recommended strawman "standard set" of such mechanisms and post it to the list for review and discussion, that would be an excellent way to energize the process. Naturally, the invitation to post such candidate contributions is extended to everyone else on the list at this time too. I hope we see some contribs in the immediate to near-term future! Cordially, BobN ------------ ISO 9001 Registered Quality Supplier ----------- Bob Natale | ACE*COMM | 301-721-3000 [v] Dir, Net Mgmt Prod | 704 Quince Orchard Rd | 301-721-3001 [f] bnatale@acecomm.com| Gaithersburg MD 20878 | www.acecomm.com ------------- Free downloads at www.winsnmp.com ------------- From owner-agentx@dorothy.bmc.com Wed Mar 1 18:37:00 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA17146 for ; Wed, 1 Mar 2000 18:36:59 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id RAA12071; Wed, 1 Mar 2000 17:29:15 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id PAA00735 for agentx-list; Wed, 1 Mar 2000 15:26:22 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id PAA00729 for ; Wed, 1 Mar 2000 15:26:17 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id RAA11556 for ; Wed, 1 Mar 2000 17:26:52 -0600 (CST) Received: (from hardaker@localhost) by wanderer.dcn.davis.ca.us (8.9.3/8.9.3) id PAA05657; Wed, 1 Mar 2000 15:26:05 -0800 To: Alexandre Briani Kieling Cc: agentx@dorothy.peer.com, ucd-snmp@ucd-snmp.ucdavis.edu Subject: Re: agentx X pure snmp References: From: Wes Hardaker X-URL: http://dcas.ucdavis.edu/~hardaker X-Microsoft: Sucks Organization: U.C.Davis, Information Technology - D.C.A.S. X-Face: #qW^}a%m*T^{A:Cp}$R\"38+d}41-Z}uU8,r%F#c#s:~Nzp0G9](s?,K49KJ]s"*7gvRgA SrAvQc4@/}L7Qc=w{)]ACO\R{LF@S{pXfojjjGg6c;q6{~C}CxC^^&~(F]`1W)%9j/iS/ IM",B1M.?{w8ckLTYD'`|kTr\i\cgY)P4 Date: 01 Mar 2000 15:26:04 -0800 In-Reply-To: Alexandre Briani Kieling's message of "Tue, 29 Feb 2000 21:39:04 -0300 (EST)" Message-ID: Lines: 24 User-Agent: Gnus/5.0804 (Gnus v5.8.4) XEmacs/21.2 (Shinjuku) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-agentx@dorothy.bmc.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list >>>>> On Tue, 29 Feb 2000 21:39:04 -0300 (EST), Alexandre Briani Kieling said: Alexandre> Is the implementation of extensible agents easier than the Alexandre> traditional way of implement snmp agents? The big win with AgentX (or any subagent protocol) is that multiple applications can attach to one running agent. With only one agent per machine, how do you manage the printing service, the ethernet devices, the web server, etc... all with one program? The only 2 answers are: 1) make them all one program (hah) 2) develop some sort of ability for them to talk to the running snmp agent. Clearly, #2 is the only reasonable choice. Alexandre> Is it true that many subagents have better performance than Alexandre> one big snmp agent? Probably not on a single processor system. On a SMP system, probably. -- Wes Hardaker Distributed Computing Analysis and Support University of California at Davis From owner-agentx@dorothy.bmc.com Thu Mar 2 05:25:21 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id FAA09457 for ; Thu, 2 Mar 2000 05:25:19 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id EAA18376; Thu, 2 Mar 2000 04:11:30 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id CAA11491 for agentx-list; Thu, 2 Mar 2000 02:09:01 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id CAA11486 for ; Thu, 2 Mar 2000 02:08:56 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id EAA18005 for ; Thu, 2 Mar 2000 04:09:31 -0600 (CST) Received: from ribble.server.csc.liv.ac.uk ([138.253.124.242]) by mailhub2.liv.ac.uk with esmtp (Exim 2.12 #1) id 12QSXh-0004Bz-00; Thu, 2 Mar 2000 10:09:13 +0000 Received: from daves.staff.csc.liv.ac.uk (root@daves.staff.csc.liv.ac.uk [138.253.124.36]) by ribble.server.csc.liv.ac.uk (8.8.8/LUCS-DTS-3.0M10) with ESMTP id KAA12972; Thu, 2 Mar 2000 10:09:13 GMT Received: from daves.staff.csc.liv.ac.uk (daves@localhost [127.0.0.1]) by daves.staff.csc.liv.ac.uk (8.8.7/LUCS-DTS-3.0D9) with ESMTP id KAA01575; Thu, 2 Mar 2000 10:08:53 GMT Message-Id: <200003021008.KAA01575@daves.staff.csc.liv.ac.uk> X-Mailer: exmh version 2.0.2 To: Alexandre Briani Kieling Cc: agentx@dorothy.peer.com, ucd-snmp@ucd-snmp.ucdavis.edu Subject: Re: agentx X pure snmp In-Reply-To: Your message of "Tue, 29 Feb 2000 21:39:04 -0300." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 02 Mar 2000 10:08:53 +0000 From: Dave Shield Sender: owner-agentx@dorothy.bmc.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list I'll add my tuppen'orth to this discussion. The main benefit of AgentX and similar technologies (as various people have said) is the additional flexibility this provides. It makes it possible to extend the agent without requiring an interruption in service. Another benefit, that hasn't been mentioned so far, is that it provides a level of security between the various subagents. A problem in one subagent, that might cause it to crash, should have no effect on the others, so the overall agent will continue to run (at a reduced level of service). In contrast, a problem in one area of a monolithic might well bring the whole thing down. Such a thing "ought not to happen", but we all know that programs have bugs in them, and ignoring the implications is foolish. > Is the implementation of extensible agents easier than > the traditional way of implement snmp agents? Hmmmm.... not sure about this. Are you talking about developing a complete agent "from scratch", or extending an agent to include support for a new MIB. Assuming the second, and a decent agent library framework, then I doubt there'd be much to choose between them. The main advantage of the subagent route would be the limited effect of errors in the new code, and not having to restart the main agent. (Though I'd typically do this on a non-standard port anyway, so as not to affect the main agent). If you're talking about developing an agent from scratch, then including subagent support is additional work over a monolithic agent, and hence harder. Even if you're talking about a "pure subagent" (which is easily the simpler side of things!), a protocol like AgentX is more complex than SNMPv1/2, so such an agent would probably be harder. My gut feeling is that SNMPv3 swings things the other way, though I haven't been involved in the UCD v3 work, so would hesitate to say this categorically. (But witness the fact that the UCD SNMPv3 work involved three or four programmers full time, while the AgentX team is li'l ol' me, mostly working of an evening or between other tasks!) > Is it true that many subagents have better performance than one big snmp > agent? Probably not. With subagents, you've immediately introduced an additional overhead for processing a request. I'd be horrified if the internal dispatch mechanism of a monolithic agent was less efficient than the AgentX handling! Even with a multi-processor system, it's probably not as simple as Wes makes out. There's nothing inherent in a monolithic agent to prevent it handling the various components of a request in parallel. The UCD agent isn't thread safe - but if it were, the architecture could be tweaked to use separate threads for each requested object without too much work. *All other things being equal*, a monolithic agent will always be faster than a subagent version. Of course, subagents provide a very simple way of parallelising the variable processing, without needing the careful design required for a multi-threaded agent. The other option open to a monolithic agent is inter-module communication. This could provide significant benefits - particularly for SET handling. AgentX (and similar mechanisms) are inherently forced to use a multi-step mechanism to achieve the "all-at-once" nature of SETs. A monolithic agent *could* potentially use direct communication to ensure that everything would be handled correctly. In fact, the UCD agent uses a multi-pass mechanism (which is convenient when it comes to the AgentX implementation), but this is not essential, in the same way as it is for subagents. As Richard points out, one performance benefit of the subagent approach is that the subagent for a particular application can be embedded in the binary for that application. This then opens up greater possibilities for efficient access to the necessary management data (which is always the hardest bit of implementing a new MIB). Essentially, you've given up possible saving in the agent, for savings in the application. This is almost invariably a Good Deal! > Is it better to develop a new module and compile it > with a existent snmp agent like ucd-snmp or is it better to develop a > extensible agent using agentX? Well, the design of the UCD agent is such that the module developer doesn't need to make that choice. The UCD Module-API is the same in both cases. I can't speak for any other development environment, but there's nothing inherent in the handling of a request that forces this distinction. > Is there any situation where only > extensible agents can do the job? Off the top of my head - any situation where the necessary management information isn't available externally to the application itself. I'll re-iterate this point, that is almost always the hardest bit of implementing a MIB module - how do you get the answers in the first place! > I do not mean to bother you with so many questions. S'OK - once I get started, it's hard to shut me up. (Whaddaya mean, you'd noticed!) Dave From owner-agentx@dorothy.bmc.com Thu Mar 2 12:50:32 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id MAA20873 for ; Thu, 2 Mar 2000 12:50:31 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id LAA13194; Thu, 2 Mar 2000 11:40:06 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id JAA12826 for agentx-list; Thu, 2 Mar 2000 09:35:09 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id JAA12821 for ; Thu, 2 Mar 2000 09:34:59 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id LAA11707 for ; Thu, 2 Mar 2000 11:35:34 -0600 (CST) Received: (from hardaker@localhost) by wanderer.dcn.davis.ca.us (8.9.3/8.9.3) id JAA14445; Thu, 2 Mar 2000 09:35:18 -0800 To: Dave Shield Cc: Alexandre Briani Kieling , agentx@dorothy.peer.com, ucd-snmp@ucd-snmp.ucdavis.edu Subject: Re: agentx X pure snmp References: <200003021008.KAA01575@daves.staff.csc.liv.ac.uk> From: Wes Hardaker X-URL: http://dcas.ucdavis.edu/~hardaker X-Microsoft: Sucks Organization: U.C.Davis, Information Technology - D.C.A.S. X-Face: #qW^}a%m*T^{A:Cp}$R\"38+d}41-Z}uU8,r%F#c#s:~Nzp0G9](s?,K49KJ]s"*7gvRgA SrAvQc4@/}L7Qc=w{)]ACO\R{LF@S{pXfojjjGg6c;q6{~C}CxC^^&~(F]`1W)%9j/iS/ IM",B1M.?{w8ckLTYD'`|kTr\i\cgY)P4 Date: 02 Mar 2000 09:35:18 -0800 In-Reply-To: Dave Shield's message of "Thu, 02 Mar 2000 10:08:53 +0000" Message-ID: Lines: 26 User-Agent: Gnus/5.0804 (Gnus v5.8.4) XEmacs/21.2 (Shinjuku) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-agentx@dorothy.bmc.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list >>>>> On Thu, 02 Mar 2000 10:08:53 +0000, Dave Shield said: Dave> Even if you're talking about a "pure subagent" (which is easily Dave> the simpler side of things!), a protocol like AgentX is more Dave> complex than SNMPv1/2, so such an agent would probably be Dave> harder. My gut feeling is that SNMPv3 swings things the other Dave> way, though I haven't been involved in the UCD v3 work, so would Dave> hesitate to say this categorically. I'd say that in general, agentx should always be easier to implement. You don't ever have to implement any form of security in the protocol. This is what made implementing an SNMPv3 implementation so much more difficult (the protocol half is easy). Dave> (But witness the fact that the UCD SNMPv3 work involved three or Dave> four programmers full time, while the AgentX team is li'l ol' Dave> me, mostly working of an evening or between other tasks!) Actually, thats not quite true... It was 3-4 part time programmers, in general. But, your single person work on agentx still shows how much easier it is to just do agentx. -- Wes Hardaker Distributed Computing Analysis and Support University of California at Davis From owner-agentx@dorothy.peer.com Mon Mar 6 08:17:24 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA03950 for ; Mon, 6 Mar 2000 08:17:24 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id HAA24994; Mon, 6 Mar 2000 07:09:34 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id FAA13687 for agentx-list; Mon, 6 Mar 2000 05:00:10 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id FAA13682 for ; Mon, 6 Mar 2000 05:00:00 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id HAA23439 for ; Mon, 6 Mar 2000 07:00:34 -0600 (CST) Received: (from uucp@localhost) by gwa2.fe.bosch.de (8.9.1/8.9.1) id OAA18746 for ; Mon, 6 Mar 2000 14:00:59 +0100 (MET) Received: from fez7540.fe.bosch.de( 10.8.2.28) by gwa2.fe.bosch.de via smap (V2.1) id xma017603; Mon, 6 Mar 00 13:59:46 +0100 Received: from leto.bk.bosch.de ([150.249.0.35]) by zeus.bk.bosch.de (Netscape Messaging Server 3.0) with ESMTP id AAA755; Mon, 6 Mar 2000 13:59:44 +0100 Received: from marconicomms.com (localhost [127.0.0.1]) by leto.bk.bosch.de (8.9.1b+Sun/8.9.1) with ESMTP id OAA23398; Mon, 6 Mar 2000 14:00:45 +0100 (MET) Message-ID: <38C3ABFD.8D0241AD@marconicomms.com> Date: Mon, 06 Mar 2000 14:00:45 +0100 From: Frank Fock Organization: Marconi Communications Software Systems GmbH & Co. KG X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.7 sun4u) X-Accept-Language: en MIME-Version: 1.0 To: Dave Shield CC: agentx@dorothy.peer.com, ucd-snmp-coders@ucd-snmp.ucdavis.edu Subject: AgentX Inter Operability Tests UCD vs. AgentX++ (First Impression) References: <200002291637.QAA00864@daves.staff.csc.liv.ac.uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Hi Dave, yesterday I did the first inter operability testings between the UCD implementation (4.1.1 on SuSE Linux 6.3) and my AgentX++ implementation. After solving a few bugs in my encoding/decoding of integers and oids I got an AgentX++ master with the agentxd subagent up and running. A walk on 0.0 works until the first OPAQUE (float) in your private UCD MIB. The debug log of agentxd shows that it gets the value from the instrumentation and then it seems that it does not send the value out on the socket. If you are interested in the debug log, I can send it to you. The test with the snmpd and an AgentX++ subagent failed. My subagents successfully register at your master, but if I do a walk on the registered subtree the snmpd debug log shows that the agentx instrumentation is called but there is no message going out on the socket. So, my original request times out. Do I need to compile the MIB defining the subtree I am registering into snmpd? I registered for example 1.3.6.1.2.1.37 (atmMIB). Today I will do some SET request and GETBULK testing. Best regards, Frank -- Frank Fock Marconi Communications Software Systems Blumenstrasse 22-24, Phone: +49 7191 13-4851 71522 Backnang, Fax: +49 7191 13-2502 Germany From owner-agentx@dorothy.peer.com Mon Mar 6 08:45:12 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id IAA05862 for ; Mon, 6 Mar 2000 08:45:12 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id HAA02385; Mon, 6 Mar 2000 07:42:46 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id FAA13753 for agentx-list; Mon, 6 Mar 2000 05:41:22 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id FAA13748 for ; Mon, 6 Mar 2000 05:41:18 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id HAA02152 for ; Mon, 6 Mar 2000 07:41:53 -0600 (CST) Received: (from uucp@localhost) by gwa2.fe.bosch.de (8.9.1/8.9.1) id OAA15901 for ; Mon, 6 Mar 2000 14:42:09 +0100 (MET) Received: from fez7540.fe.bosch.de( 10.8.2.28) by gwa2.fe.bosch.de via smap (V2.1) id xma015419; Mon, 6 Mar 00 14:41:30 +0100 Received: from leto.bk.bosch.de ([150.249.0.35]) by zeus.bk.bosch.de (Netscape Messaging Server 3.0) with ESMTP id AAA4541 for ; Mon, 6 Mar 2000 14:41:27 +0100 Received: from marconicomms.com (localhost [127.0.0.1]) by leto.bk.bosch.de (8.9.1b+Sun/8.9.1) with ESMTP id OAA23413 for ; Mon, 6 Mar 2000 14:42:27 +0100 (MET) Message-ID: <38C3B5C3.2134F8C6@marconicomms.com> Date: Mon, 06 Mar 2000 14:42:27 +0100 From: Frank Fock Organization: Marconi Communications Software Systems GmbH & Co. KG X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.7 sun4u) X-Accept-Language: en MIME-Version: 1.0 To: agentx Subject: snmpDomain OID for UNIX domain sockets? Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Hi, if I have not overlooked something, there is no snmpDomain OID for UNIX domain sockets defined in the SNMPv2-TM MIB nor in the AGENTX-MIB. So, what OID should I use when setting up an agentxConnTransportDomain object for an AgentX UNIX domain socket connection? Where is it defined, if it is defined somewhere? TIA, Frank Fock -- Frank Fock Marconi Communications Software Systems Blumenstrasse 22-24, Phone: +49 7191 13-4851 71522 Backnang, Fax: +49 7191 13-2502 Germany From owner-agentx@dorothy.peer.com Mon Mar 6 14:22:15 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA19799 for ; Mon, 6 Mar 2000 14:22:14 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id NAA09842; Mon, 6 Mar 2000 13:17:31 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id LAA15790 for agentx-list; Mon, 6 Mar 2000 11:15:07 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id LAA15785 for ; Mon, 6 Mar 2000 11:15:00 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id NAA09254 for ; Mon, 6 Mar 2000 13:15:35 -0600 (CST) Received: (from hardaker@localhost) by wanderer.dcn.davis.ca.us (8.9.3/8.9.3) id LAA16867; Mon, 6 Mar 2000 11:14:26 -0800 To: Frank Fock Cc: Dave Shield , agentx@dorothy.peer.com, ucd-snmp-coders@ucd-snmp.ucdavis.edu Subject: Re: AgentX Inter Operability Tests UCD vs. AgentX++ (First Impression) References: <200002291637.QAA00864@daves.staff.csc.liv.ac.uk> <38C3ABFD.8D0241AD@marconicomms.com> From: Wes Hardaker X-URL: http://dcas.ucdavis.edu/~hardaker X-Microsoft: Sucks Organization: U.C.Davis, Information Technology - D.C.A.S. X-Face: #qW^}a%m*T^{A:Cp}$R\"38+d}41-Z}uU8,r%F#c#s:~Nzp0G9](s?,K49KJ]s"*7gvRgA SrAvQc4@/}L7Qc=w{)]ACO\R{LF@S{pXfojjjGg6c;q6{~C}CxC^^&~(F]`1W)%9j/iS/ IM",B1M.?{w8ckLTYD'`|kTr\i\cgY)P4 Date: 06 Mar 2000 11:14:26 -0800 In-Reply-To: Frank Fock's message of "Mon, 06 Mar 2000 14:00:45 +0100" Message-ID: Lines: 13 User-Agent: Gnus/5.0804 (Gnus v5.8.4) XEmacs/21.2 (Shinjuku) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list >>>>> On Mon, 06 Mar 2000 14:00:45 +0100, Frank Fock said: Frank> Do I need to compile the MIB defining the subtree I am Frank> registering into snmpd? I registered for example 1.3.6.1.2.1.37 Frank> (atmMIB). I'll let Dave respond to the rest when he gets back from his long weekend, but no you do not need to do the above. -- Wes Hardaker Distributed Computing Analysis and Support University of California at Davis From owner-agentx@dorothy.peer.com Mon Mar 6 23:44:49 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id XAA24172 for ; Mon, 6 Mar 2000 23:44:49 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id WAA23192; Mon, 6 Mar 2000 22:43:15 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id UAA01155 for agentx-list; Mon, 6 Mar 2000 20:39:50 -0800 (PST) Received: (from rpresuhn@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id UAA01114 for agentx@dorothy.bmc.com; Mon, 6 Mar 2000 20:39:32 -0800 (PST) Date: Mon, 6 Mar 2000 20:39:32 -0800 (PST) From: Randy Presuhn Message-Id: <200003070439.UAA01114@dorothy.bmc.com> To: agentx@dorothy.peer.com Subject: Fwd: non-subscriber posting to agentx Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Hi - A posting to the agentx list from a non-subscriber address... This address has been added to the "posters" list. ------------------------------------------------------------------------ Randy Presuhn randy_presuhn@bmc.com http://www.bmc.com/ Voice: +1 408 546-1006 BMC Software, Inc. 1-3141 2141 N. First Street Fax: +1 408 965-0359 San Jose, California 95131 USA ------------------------------------------------------------------------ Any relationship between my opinions and BMC's should be coincidental. ------------------------------------------------------------------------ > Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) > by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id SAA27780 > for ; Mon, 6 Mar 2000 18:05:31 -0800 (PST) > Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) > by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id UAA01601 > for ; Mon, 6 Mar 2000 20:06:07 -0600 (CST) > Received: from mira1.cisco.com (mira1.cisco.com [171.71.212.196]) > by kickme.cisco.com (8.9.1a/8.9.1) with ESMTP id RAA28521; > Mon, 6 Mar 2000 17:54:38 -0800 (PST) > Received: from cisco.com (dhcp-171-71-213-186.cisco.com [171.71.213.186]) > by Mirapoint Server mira1.cisco.com > with ESMTP id ABV38207 > Mon, 6 Mar 2000 18:06:02 -0800 (PST) > Message-ID: <38C462F1.E9701A39@cisco.com> > Date: Mon, 06 Mar 2000 18:01:21 -0800 > From: Lauren Heintz > Organization: Cisco Systems > X-Mailer: Mozilla 4.5 [en] (WinNT; U) > X-Accept-Language: en > MIME-Version: 1.0 > To: Frank Fock > CC: agentx > Subject: Re: snmpDomain OID for UNIX domain sockets? > References: <38C3B5C3.2134F8C6@marconicomms.com> > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: 7bit > > Hi, > > My recollection is that we avoided putting these kinds of domain defs in the > AgentX > MIB because they were to be made std in a specialized MIB or via IANA, > etc.. In the event a std def is lacking, you could always define a > proprietary TDomain value and implement around that, and as std defs become > avaialable, support those as well (or instead). > > Regards, > Lauren > > Frank Fock wrote: > > > Hi, > > > > if I have not overlooked something, there is no snmpDomain OID for UNIX > > domain sockets defined in the SNMPv2-TM MIB nor in the AGENTX-MIB. > > > > So, what OID should I use when setting up an agentxConnTransportDomain > > object for an AgentX UNIX domain socket connection? Where is it defined, > > if it is defined somewhere? > > > > TIA, > > Frank Fock > > > > -- > > Frank Fock Marconi Communications Software Systems > > Blumenstrasse 22-24, > > Phone: +49 7191 13-4851 71522 Backnang, > > Fax: +49 7191 13-2502 Germany > > From owner-agentx@dorothy.peer.com Tue Mar 7 06:49:34 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA03858 for ; Tue, 7 Mar 2000 06:49:33 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id FAA25958; Tue, 7 Mar 2000 05:47:33 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id DAA09277 for agentx-list; Tue, 7 Mar 2000 03:45:12 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id DAA09272 for ; Tue, 7 Mar 2000 03:45:07 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id FAA25693 for ; Tue, 7 Mar 2000 05:45:42 -0600 (CST) Received: from henkell.ibr.cs.tu-bs.de (schoenw@henkell [134.169.34.191]) by mumm.ibr.cs.tu-bs.de (8.9.3/8.9.3) with ESMTP id MAA07641; Tue, 7 Mar 2000 12:45:40 +0100 (MET) Received: from schoenw@localhost by henkell.ibr.cs.tu-bs.de (8.7.6/tubsibr) id MAA06745; Tue, 7 Mar 2000 12:45:39 +0100 Date: Tue, 7 Mar 2000 12:45:39 +0100 Message-Id: <200003071145.MAA06745@henkell.ibr.cs.tu-bs.de> From: Juergen Schoenwaelder To: frank.fock@marconicomms.com CC: agentx@dorothy.peer.com In-reply-to: <38C3B5C3.2134F8C6@marconicomms.com> (message from Frank Fock on Mon, 06 Mar 2000 14:42:27 +0100) Subject: Re: snmpDomain OID for UNIX domain sockets? References: <38C3B5C3.2134F8C6@marconicomms.com> Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list >>>>> Frank Fock writes: Frank> if I have not overlooked something, there is no snmpDomain OID Frank> for UNIX domain sockets defined in the SNMPv2-TM MIB nor in the Frank> AGENTX-MIB. Correct. Frank> So, what OID should I use when setting up an Frank> agentxConnTransportDomain object for an AgentX UNIX domain Frank> socket connection? Where is it defined, if it is defined Frank> somewhere? The AGENTX-MIB is an example showing the need for more concrete TAddress definitions. There is also a desire in other WGs to use a centrally administered enumeration rather than OIDs to identify address types. The IPv6-MIB design team is motivated to write a document which makes IANA responsible for maintaining address type assignments. This document will also provide additional concrete TAddress definitions. But there is no concrete document out there yet since we first want to get the INET-ADDRESS-MIB through IETF last call and IESG review. Once such a document is in place, the AgentX WG has to decide whether they want to adopt it for the AGENTX-MIB or not (which probably means to recycle once at Proposed). /js -- Juergen Schoenwaelder Technical University Braunschweig Dept. Operating Systems & Computer Networks Phone: +49 531 391 3289 Bueltenweg 74/75, 38106 Braunschweig, Germany Fax: +49 531 391 5936 From owner-agentx@dorothy.peer.com Tue Mar 7 09:58:28 2000 Received: from tangelo.bmc.com (fw-us-hou-1.bmc.com [198.207.223.250]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id JAA16941 for ; Tue, 7 Mar 2000 09:58:27 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id IAA22991; Tue, 7 Mar 2000 08:50:43 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id GAA09640 for agentx-list; Tue, 7 Mar 2000 06:44:40 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id GAA09635 for ; Tue, 7 Mar 2000 06:44:36 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id IAA20909 for ; Tue, 7 Mar 2000 08:45:11 -0600 (CST) Received: by mailext04.compaq.com (Postfix, from userid 12345) id C9517104C53; Tue, 7 Mar 2000 08:45:02 -0600 (CST) Received: from mailint12.im.hou.compaq.com (mailint12.compaq.com [207.18.199.190]) by mailext04.compaq.com (Postfix) with ESMTP id C50F3FB101; Tue, 7 Mar 2000 08:45:02 -0600 (CST) Received: by mailint12.im.hou.compaq.com (Postfix, from userid 12345) id E7BB34FB06; Tue, 7 Mar 2000 08:44:55 -0600 (CST) Received: from sigma.zk3.dec.com (brysigma.zk3.dec.com [16.141.40.6]) by mailint12.im.hou.compaq.com (Postfix) with ESMTP id 888434C904; Tue, 7 Mar 2000 08:44:55 -0600 (CST) Received: from zk3.dec.com by sigma.zk3.dec.com (8.8.8/1.1.20.3/24Apr98-0811AM) id JAA0000007084; Tue, 7 Mar 2000 09:45:01 -0500 (EST) Message-ID: <38C51505.E8DFD8B1@zk3.dec.com> Date: Tue, 07 Mar 2000 09:41:09 -0500 From: Mike Daniele Organization: Compaq 64-bit UNIX networking X-Mailer: Mozilla 4.72 [en] (X11; I; OSF1 V5.0 alpha) X-Accept-Language: en MIME-Version: 1.0 To: Frank Fock Cc: agentx Subject: Re: snmpDomain OID for UNIX domain sockets? References: <38C3B5C3.2134F8C6@marconicomms.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Hi, As Lauren's reply said, we didn't feel the AgentX MIB was the place to define TDomain/TAddress values. We felt (as do many) there should be a single document with such defs, because they are really independent of application protocol. And in fact, now that we're just about done defining inet address constructs (draft-ops-endpoint-mib-07.txt), I believe we'll start on a draft for transport constructs. Mike Frank Fock wrote: > Hi, > > if I have not overlooked something, there is no snmpDomain OID for UNIX > domain sockets defined in the SNMPv2-TM MIB nor in the AGENTX-MIB. > > So, what OID should I use when setting up an agentxConnTransportDomain > object for an AgentX UNIX domain socket connection? Where is it defined, > if it is defined somewhere? > > TIA, > Frank Fock > > -- > Frank Fock Marconi Communications Software Systems > Blumenstrasse 22-24, > Phone: +49 7191 13-4851 71522 Backnang, > Fax: +49 7191 13-2502 Germany From owner-agentx@dorothy.peer.com Tue Mar 7 13:13:39 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id NAA23187 for ; Tue, 7 Mar 2000 13:13:38 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id MAA01664; Tue, 7 Mar 2000 12:06:01 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id KAA21679 for agentx-list; Tue, 7 Mar 2000 10:02:52 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id KAA20602 for ; Tue, 7 Mar 2000 10:02:01 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id MAA00665 for ; Tue, 7 Mar 2000 12:02:35 -0600 (CST) Received: (from uucp@localhost) by gwa2.fe.bosch.de (8.9.1/8.9.1) id TAA16013 for ; Tue, 7 Mar 2000 19:02:16 +0100 (MET) Received: from fez7540.fe.bosch.de( 10.8.2.28) by gwa2.fe.bosch.de via smap (V2.1) id xma015919; Tue, 7 Mar 00 19:02:07 +0100 Received: from leto.bk.bosch.de ([150.249.0.35]) by zeus.bk.bosch.de (Netscape Messaging Server 3.0) with ESMTP id AAA29212; Tue, 7 Mar 2000 19:02:09 +0100 Received: from marconicomms.com (localhost [127.0.0.1]) by leto.bk.bosch.de (8.9.1b+Sun/8.9.1) with ESMTP id TAA24167; Tue, 7 Mar 2000 19:03:06 +0100 (MET) Message-ID: <38C54459.5340A7CE@marconicomms.com> Date: Tue, 07 Mar 2000 19:03:05 +0100 From: Frank Fock Organization: Marconi Communications Software Systems GmbH & Co. KG X-Mailer: Mozilla 4.7 [en] (X11; I; SunOS 5.7 sun4u) X-Accept-Language: en MIME-Version: 1.0 To: Dave Shield , agentx@dorothy.peer.com, ucd-snmp-coders@ucd-snmp.ucdavis.edu Subject: Re: AgentX Inter Operability Tests UCD vs. AgentX++ (First Impression) References: <200002291637.QAA00864@daves.staff.csc.liv.ac.uk> <38C3ABFD.8D0241AD@marconicomms.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Frank Fock wrote: > > The test with the snmpd and an AgentX++ subagent failed. My subagents > successfully register at your master, but if I do a walk on the > registered subtree the snmpd debug log shows that the agentx instrumentation > is called but there is no message going out on the socket. So, my > original request times out. > I have tested the UCD master / AgentX++ subagent combination once again. The statement above from my first mail is not correct. My subagent receives AgentX search requests but the upper bound is set to the lower bound of the registered region by the UCD master. Because of this, my subagent is returning endOfMibView. The master seems to not make a error status conversion from v2c to v1. It returns on the GETNEXT request. The region I am registering is the atmMIB (1.3.6.1.2.1.37). The GETNEXT I issued was on 1.3.6.1.2.1.37. > Today I will do some SET request and GETBULK testing. > The SET request processings I have done so far worked. But I will have to create some more complex test cases. The GETBULK work fine with the UCD master and AgentX++ subagent, except the problem with OPAQUE types mentioned in my first mail. Best regards, Frank -- Frank Fock Marconi Communications Software Systems Blumenstrasse 22-24, Phone: +49 7191 13-4851 71522 Backnang, Fax: +49 7191 13-2502 Germany From owner-agentx@dorothy.peer.com Wed Mar 8 06:17:04 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id GAA24162 for ; Wed, 8 Mar 2000 06:17:03 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id FAA24195; Wed, 8 Mar 2000 05:14:33 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id DAA06629 for agentx-list; Wed, 8 Mar 2000 03:11:11 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id DAA06624 for ; Wed, 8 Mar 2000 03:11:05 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id FAA23833 for ; Wed, 8 Mar 2000 05:11:41 -0600 (CST) Received: from ribble.server.csc.liv.ac.uk ([138.253.124.242]) by mailhub1.liv.ac.uk with esmtp (Exim 2.12 #1) id 12SeLt-0004aH-00; Wed, 8 Mar 2000 11:10:05 +0000 Received: from daves.staff.csc.liv.ac.uk (root@daves.staff.csc.liv.ac.uk [138.253.124.36]) by ribble.server.csc.liv.ac.uk (8.8.8/LUCS-DTS-3.0M10) with ESMTP id LAA06336; Wed, 8 Mar 2000 11:10:04 GMT Received: from daves.staff.csc.liv.ac.uk (daves@localhost [127.0.0.1]) by daves.staff.csc.liv.ac.uk (8.8.7/LUCS-DTS-3.0D9) with ESMTP id LAA16677; Wed, 8 Mar 2000 11:10:17 GMT Message-Id: <200003081110.LAA16677@daves.staff.csc.liv.ac.uk> X-Mailer: exmh version 2.0.2 To: Frank Fock To: agentx@dorothy.peer.com, ucd-snmp-coders@ucd-snmp.ucdavis.edu Subject: Re: AgentX Inter Operability Tests UCD vs. AgentX++ (First Impression) In-Reply-To: Your message of "Tue, 07 Mar 2000 19:03:05 +0100." <38C54459.5340A7CE@marconicomms.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 08 Mar 2000 11:10:16 +0000 From: Dave Shield Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list > I have tested the UCD master / AgentX++ subagent combination once > again. The statement above from my first mail is not correct. My > subagent receives AgentX search requests but the upper bound is > set to the lower bound of the registered region by the UCD master. Oh, that's a relief. I knew that the UCD agent didn't handle ranges correctly, but it was a bit worrying that it wasn't sending out the request at all. Can you try the following patch (untested, but it should work :-) ) -------------- snip ---------- *** master_request.c 2000/03/08 10:52:34 1.1 --- master_request.c 2000/03/08 10:54:16 *************** *** 339,345 **** sub = find_subtree_previous( vbp->name, vbp->name_length, NULL ); snmp_pdu_add_variable( request->pdu, vbp->name, vbp->name_length, ASN_PRIV_INCL_RANGE, ! (u_char*)sub->name, sub->namelen); } return AGENTX_ERR_NOERROR; --- 339,345 ---- sub = find_subtree_previous( vbp->name, vbp->name_length, NULL ); snmp_pdu_add_variable( request->pdu, vbp->name, vbp->name_length, ASN_PRIV_INCL_RANGE, ! (u_char*)sub->end, sub->end_len); } return AGENTX_ERR_NOERROR; -------------- snip ---------- > Because of this, my subagent is returning endOfMibView. The master > seems to not make a error status conversion from v2c to v1. Yes - that's correct There's a comment in the 'handle_agentx_response' routine to remind me that this needs fixing. > The region I am registering is the atmMIB (1.3.6.1.2.1.37). The > GETNEXT I issued was on 1.3.6.1.2.1.37. The master ought to be picking up any following (internal) modules, so this will should only bite if the last subagent registration follows any internal registrations. How did you configure the UCD master? Does it still include any of the UCD-specific support (e.g. ucd-snmp/versioninfo) ? > The SET request processings I have done so far worked. But I will > have to create some more complex test cases. Note that it's likely that failed subagent SETs will probably not be handled correctly. While I believe this worked initially, subsequent changes to the main agent (unrelated to AgentX) have broken the handling of errors from sub-agents, and this wasn't spotted until after the 4.1.1 release. > The GETBULK work fine with the UCD master and AgentX++ subagent, except > the problem with OPAQUE types mentioned in my first mail. Hmmm.... I'm slightly puzzled about this. It should be processed in precisely the same way as a string or IP Address type. I presume these are working OK? If you can send me the log, I'll see if I can spot the problem. Dave From owner-agentx@dorothy.peer.com Wed Mar 15 16:48:36 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id QAA16122 for ; Wed, 15 Mar 2000 16:48:34 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id PAA23809; Wed, 15 Mar 2000 15:43:50 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id NAA01388 for agentx-list; Wed, 15 Mar 2000 13:37:30 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id NAA01383 for ; Wed, 15 Mar 2000 13:37:26 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id PAA21975 for ; Wed, 15 Mar 2000 15:38:02 -0600 (CST) Received: by ORMAIL2 with Internet Mail Service (5.5.2650.21) id ; Wed, 15 Mar 2000 23:37:25 +0200 Message-ID: From: Dror Bar-Lev To: "'agentx@dorothy.bmc.com'" Subject: AgentX Date: Wed, 15 Mar 2000 23:37:24 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-Type: text/plain; charset="iso-8859-8" Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Hello, My name is Dror Barlev, a project manager of the NMS at Orckit which is a leading provider of Digital Subscriber Line (DSL) Solutions. We are required to implement a topology of several tens of network elements in a single SNMP/IP element. each one of the network elements is currently implemented with a SNMP proxy agent (using Epilogue and pSOS 2.5), supporting a various standard MIBs (in SNMP-V2) and few other proprietaries. These Agents should be transparent to the SNMP manager and should be presented as a single SNMP agent. We are looking for consulting on AgentX system design and implementation. I'll be appreciate if you could help us with a point of contact. Best Regards, Dror Barlev Project Manager - NMS Orckit Communications Ltd. 38 Nahalat Yitzhak St. Tel Aviv 67448, Israel Tel: +972-3-6945343 Fax: +972 3 6094754 Web Site: http://www.orckit.com E-mail: drorbl@orckit.com From owner-agentx@dorothy.peer.com Tue Mar 28 22:38:45 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id WAA03612 for ; Tue, 28 Mar 2000 22:38:44 -0500 (EST) Received: from dorothy.peer.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id VAA03319; Tue, 28 Mar 2000 21:36:15 -0600 (CST) Received: (from root@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id TAA20220 for agentx-list; Tue, 28 Mar 2000 19:29:27 -0800 (PST) Received: (from rpresuhn@localhost) by dorothy.peer.com (8.8.6 (PHNE_12836)/8.8.6) id TAA20201; Tue, 28 Mar 2000 19:29:18 -0800 (PST) Date: Tue, 28 Mar 2000 19:29:18 -0800 (PST) From: Randy Presuhn Message-Id: <200003290329.TAA20201@dorothy.peer.com> To: agentx@dorothy.peer.com Subject: fwd: non-subscriber post to agent list Cc: shivendra.kumar@wipro.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit Hi - a non-subscriber post to the agent list. ------------------------------------------------------------------------ Randy Presuhn randy_presuhn@bmc.com http://www.bmc.com/ Voice: +1 408 546-1006 BMC Software, Inc. 1-3141 2141 N. First Street Fax: +1 408 965-0359 San Jose, California 95131 USA ------------------------------------------------------------------------ Any relationship between my opinions and BMC's should be coincidental. ------------------------------------------------------------------------ > Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) > by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id JAA16444 > for ; Tue, 28 Mar 2000 09:42:00 -0800 (PST) > Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) > by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id LAA22140 > for ; Tue, 28 Mar 2000 11:42:37 -0600 (CST) > Received: from ecvwall1.wipro.com (ecvwall1.wipro.com [192.168.181.23]) > by wiproecmx2.wipro.com (8.9.3/8.9.3) with SMTP id XAA26226 > for ; Tue, 28 Mar 2000 23:20:03 GMT > Received: from wipro.com ([192.168.178.17]) by ecmail.wipsys.soft.net > (Netscape Messaging Server 3.6) with ESMTP id AAA12CF; > Tue, 28 Mar 2000 23:06:38 +0530 > Message-ID: <38E0EF9C.AB144F8B@wipro.com> > Date: Tue, 28 Mar 2000 23:15:00 +0530 > From: "Shivendra Kumar" > Reply-To: shivendra.kumar@wipro.com > Organization: Wipro Technologies > X-Mailer: Mozilla 4.6 [en] (WinNT; I) > X-Accept-Language: en > MIME-Version: 1.0 > To: bwijnen@lucent.com > CC: agentx@dorothy.peer.com, Sanjay Gera , > Shivendra Kumar , > Abhishek Bagchi , > Amit Srivastava , > Sarath Babu Govindarajulu , randy@psg.com > Subject: More Error codes on SNMP V1? > Content-Type: multipart/mixed; > boundary="------------B24D4DC2C9EF6E85F9F3039D" > > This is a multi-part message in MIME format. > > --------------B24D4DC2C9EF6E85F9F3039D > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: 7bit > > Hi all, > > We are working on SNMP interfaces for a product named Flash600 ( ADX, > Layer2 switching, ATM) > for FNC Inc. > We are using SNMP V1 framework , but, unfortunately SNMP v1 supports > only > following error codes: > > noError(0): no error in the requested PDU. > toobig(1): The get-response message is bigger than that the local > implementation can handle. > noSuchName(2): one of the requested objects does not match anything in > the relevant MIB view that can be returned. > badValue(3): The set-request asked the agent to write an inappropriate > value. > readOnly(4): A set-request tried to write a value that the operator is > not allowed to write.Either the access specified is > READ-ONLY or the the variable MIB definition does not permit write > access. > genErr(5): A variable cannot be retrieved for reasons outside the ones > listed above. > > This provides very little granularity for the User to decide what went > wrong. > > IS THERE ANY WAY WE CAN ADD MORE ERROR CODES WITHOUT breaking V1 > compliance? > We don't want to move over to SNMPv2 ,but, still want to add more > error codes? > Can we add more error codes? If so, how? > > regards, > shivendra > > > -- > ------------------------------------------------------------------------ > > Shivendra Kumar > 56, 4th Cross, Wipro Infotech- Telecom Sols. > 5th Block, 72, Electronics City > Koramanagala, Bangalore > Bangalore -95 > mobile-98440-92528 hello(off) -0091-80-8520408/416 > > extn.2107 > fax- 0091-80-8520478 > mail@ - shivendra.kumar@wipro.com > shivendra_kumar@hotmail.com > ------------------------------------------------------------------------ > > My three warriors - perseverance, smile and hope. > > ------------------------------------------------------------------------ > > > > --------------B24D4DC2C9EF6E85F9F3039D > Content-Type: text/x-vcard; charset=us-ascii; > name="shivendra.kumar.vcf" > Content-Transfer-Encoding: 7bit > Content-Description: Card for Shivendra Kumar > Content-Disposition: attachment; > filename="shivendra.kumar.vcf" > > begin:vcard > n:Kumar;Shivendra > tel;cell:98440-92528 > tel;work:0091-80-8520420 extn 2107 > x-mozilla-html:FALSE > org:Wipro Technologies;Telecom Solutions > version:2.1 > email;internet:shivendra.kumar@wipro.com > title:Systems Manager > adr;quoted-printable:;;72, Keonics Electronics city =0D=0AHosur Main Road,=0D=0ABangalore- 561229;Bangalore;;;India > fn:Shivendra Kumar > end:vcard > > --------------B24D4DC2C9EF6E85F9F3039D-- > > From owner-agentx@dorothy.peer.com Wed Mar 29 00:50:40 2000 Received: from tangelo.bmc.com (fw-us-hou-2.bmc.com [198.207.223.251]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id AAA05156 for ; Wed, 29 Mar 2000 00:50:39 -0500 (EST) Received: from dorothy.bmc.com (dorothy.bmc.com [192.146.153.65]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with ESMTP id XAA22259; Tue, 28 Mar 2000 23:47:20 -0600 (CST) Received: (from root@localhost) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) id VAA21686 for agentx-list; Tue, 28 Mar 2000 21:43:51 -0800 (PST) Received: from tangelo.bmc.com (root@tangelo [172.17.7.166]) by dorothy.bmc.com (8.8.6 (PHNE_12836)/8.8.6) with ESMTP id VAA21681 for ; Tue, 28 Mar 2000 21:43:46 -0800 (PST) Received: from fw-us-hou1.bmc.com (fw-us-hou1.bmc.com [172.17.0.250]) by tangelo.bmc.com (8.8.6 (PHNE_17135)/8.8.6) with SMTP id XAA21802 for ; Tue, 28 Mar 2000 23:44:23 -0600 (CST) Received: from tmax (dial3-45.nbn.net [208.139.67.135]) by mail.nbn.net (8.9.3/8.9.3) with SMTP id AAA16295; Wed, 29 Mar 2000 00:44:15 -0500 Reply-To: From: "Max Devlin" To: , Subject: RE: non-subscriber post to agent list Date: Wed, 29 Mar 2000 00:44:03 -0500 Message-ID: <001101bf9941$cae3e480$87438bd0@tmax.eltrax.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0 Importance: Normal In-Reply-To: <200003290329.TAA20201@dorothy.peer.com> X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Sender: owner-agentx@dorothy.peer.com Precedence: bulk List-Id: IETF Agentx Working Group mailing list Content-Transfer-Encoding: 7bit > > We are working on SNMP interfaces for a product named > Flash600 ( ADX,[...] > > genErr(5): A variable cannot be retrieved for reasons > > outside the ones listed above. > > > > This provides very little granularity for the User to > > decide what went wrong. > > IS THERE ANY WAY WE CAN ADD MORE ERROR CODES WITHOUT > > breaking V1 compliance? > > We don't want to move over to SNMPv2 ,but, still want > > to add more error codes? > > Can we add more error codes? If so, how? > > > > regards, > > shivendra Well, since this came in late, doesn't really relate to agentx, and nobody's heard from me in a while, I thought I'd give it a shot. Such a "pre-emptive" response does run the risk of illustrating that I don't know what I'm talking about, since I'm not a programmer or engineer. But my specialty is "end user operational functionality", and it sounds like there's a chance I might be helpful, so I'll try. As indicated by my edits from the original message, I think the standard approach would be: a) have your agent return genErr b) define a MIB group to identify the specific error c) have your manager query your errorMib objects when it gets a genErr to determine what the problem is I don't know if anyone's already worked up something like this in practice. I believe that given a switch element manager you will probably have special requirements, even though the ATM stuff might lend itself to a standard effort. AFAIK, you'd be breaking new ground, but if you are planning on sticking with v1, it doesn't sound very fertile. I HIGHLY recommend getting some advice from Dave Perkins and/or other MIB developers if you go this route. (I don't mean to be "donating" anyone's time; it would be well worth the expense if you can pay for such advice, though there is enough reason for the community to desire "well behaved" implementations that you may not need to.) This kind of MIB can get much trickier than you might think up front, and I'm not at all sure I could even begin to consider all the potential issues which would arise. Some possible ideas that spring to mind, however: * Could you use the request ID from the original query as an index for a table of errors? * You could either create a table to configure the errorMib table (how big, age or wrap info, etc.) * If you feel comfortable with a limited solution, you could probably simply create a single attribute, using an enumerated integer for your list of additional error codes. * This "single object" could be a scalar if you think your manager/agent are "welded together" enough to do requests/responses in an almost synchronous manner, if you know what I mean. This would certainly be a rather dirty design, and would create a pseudo-proprietary agent/manager scenario, but you wouldn't be the first. (Don't shoot me; I'm just the piano player.) * If this issue comes up a lot, you might even program your manager to always include the OID for the error object(s) in appropriate queries. That way you can have the agent return the error information along with the original response without breaking v1's rule about the response varbinds being the same as the request's. * A complete approach would probably include timestamps, manager identity (community, IP), and the OID which generated the problem. * Come to think of it, unless you use single varbind requests, you'll HAVE to reference the OID, though you might still do it "the easy way" and avoid using a table. Just define two objects, one for the OID which 'didn't work' (I'm thinking you're considering SETs more than GETs, though according to the mechanics of SNMP, it really doesn't matter) and one for the error code. If you use a table, you could index by request ID (if that's even feasible) and OID. When the manager gets a genErr, it "looks it up" by querying with a mib walk using your object.requestID.0, and it will retrieve all the OIDs corresponding to the varbind(s) that caused errors. If you find that I'm full of hot air, and somebody else already *has* worked this stuff out (or I'm just missing something), could you let me know? Thanks for your time. Hope it helps.