- db-wg - mailman.ripe.net

Re: db-wg digest, Vol 1 #152 - 13 msgs
by Rev Adrian Kennard 12 Jan '04

12 Jan '04

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 | From: Thor Kottelin <thor(a)anta.net> | So, we already have the irt object and the trouble attribute. | Additionally, | abuse is often reported based on admin-c, tech-c, changed, remarks and | $DEITY knows what. Please let's not add to the confusion by creating | even more attribute or object types. There is a clear problem with what we have now, whether techncially or simply awareness, as it is clear that lots and lots of inet[6]num objects have remarks identifying abuse contacts in a variety of formats. There is a need for people to identify abuse contacts associated with an IP[6]. "trouble" does not help as it is also free format and so not much better than remarks. "irt" may be the answer, and I am creating one for our use at present (still not had a reply - why is it a manual process?). It may be over complex. It is not well publicised to those that need to know. Even those that know of it (and I had seen it) do not realise it is the way to put an abuse contact on an inetnum (hence my original question). FYI, our abuse report scripts now look for a mnt-irt field and if not found go for the admin-c field. They used to just go for the admin field. However, I do think something has to be done as it is a nonsense the way it is now. - -- Rev Adrian Kennard -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFAAufMHBb4e52L0Y0RAoHlAJ0ZLrp5QiftL7iWEqpcUxVUxZjLSgCcC8Vo j86xkD4uQ0dcTkzcWOOjtaU= =ykiB -----END PGP SIGNATURE-----

1 0

Re: [db-wg] abuse-c
by Wilfried Woeber, UniVie/ACOnet 12 Jan '04

12 Jan '04

Hi John! >Menno Pieters (Stelvio) wrote: >> TI is an organisation that verifies the information, so that it can be >> *trusted*. For ordinary users it may be to hard to understand this and >> how to make sure it all correct. For IRT among eachother it may be >> important. > >But all that going via the TI route gives you is a "mnt-by: TI" entry in >your object. Yes - what is wrong with that? > There are other non-RIPE ways for teams to infer how much >trust they should give to another team, Correct. >and overloading mnt-by seems undesirable. Maybe, but I don't see what the downside is in using a mechanism which is there, instead of inventing something new. Actually it was - maybe - blue-eyed to invent a new object type. Inventing a new attribute _without a proper syntax and semantic to go with it_ is probably not too much better... >Some useful changes would be renaming the IRT object to something more >general, like Abuse (?) object If we can reach consensus that this should be the name I expect the NCC to make the necessary modifications to the implementation. The problem (which was discussed in various environments more than once!) is that different entities have a different interpretation of the word "abuse". >and for it to be returned by default when querying an IP (or AS). I keep hearing that again and again (and from the incident/abuse point of view I agree!), so I'll put it onto the DB-WG. Maybe others can try to put it onto other WGs' agendas, as this is a pretty big change to the behaviour of the whois server. >The seperation into an easily maintainable object which includes stuff >in addition to email address (like postal address, pgp and phone) are >useful enhancements to "trouble:". > >Cheers >John Wilfried.

2 1

Re: [db-wg] abuse-c
by Wilfried Woeber, UniVie/ACOnet 12 Jan '04

12 Jan '04

> Not if the relevant IRT object is automatically returned for every > inetnum query. This is what ARIN does, BTW. Could I see an example of ARIN doing that, please? Or - what is the query to get this from ARIN? Thanks, Wilfried.

2 1

RE: [db-wg] abuse-c
by Wilfried Woeber, UniVie/ACOnet 12 Jan '04

12 Jan '04

Christian, >I have a lot of inetnum objects which I need the same abuse address for so I >would REALLY appriciate not having to change each and everyone of them!.. :) supporting that was one of the goals in the design of the IRT object. All the (resources described by a set of) objects which are handled by the same incident team or function just get linked to the same IRT object. If you want to make any changes, then you update the info in _one_ place - much like the maintainer approach. And, the IRT structure supports hierarchy (for those resources which do have a hierarchy :-) It allows quite a few other things to be nicely described and maintained, eg. - giving key info if you want to use encryption for your complaint, or verify the identity of the security contact you get an answer from - decoupling of IP-Address management from operations ("rouitng") and incident handling - registering credibility info with this contact info, like the Trusted Introducer, which was the first structure to use this machinery. We are definitely still missing some things to make it fly high... Things that come to my mind without doing a full inventory: - make additional documentation available (TF-CSIRT is actively working on this) - try to get it deployed and _used_ (I'd look into the direction of the Anti-Spam folks to help with this) - maybe modify the default answer from the server for a whois query (what would the impact be on the "other" users of the registry data?) - try to define complaint handling roles in a concise way to allow selecting the "correct" contact from an object. (this doesn't chance whether you put it into an inetnum or an IRT object ;-) (I guess there will be another attempt in Madrid next week to deal with that amongst the incident handlers, and we may come back with a proposal or at least an idea) Wilfried.

1 0

RE: [db-wg] abuse-c
by TAYON, Julien 12 Jan '04

12 Jan '04

John Green <mailto:j.green@ukerna.ac.uk> wrote: > I have never understood what this gives you. If "Evil Company" wants > to misdirect abuse reports (why?) they can circumvent this by making > a fake IRT object with IRT XYZ as the contact email address. > If people circumvent this they are either evil or incompetent which is alreday a strong hint to be suspicious. Julien Tayon LDCOM networks BU DATA- +33 1 70 18 14 45 1 square Chaptal , 92309 Levallois Cedex

4 4

Defined roles
by Rev Adrian Kennard 12 Jan '04

12 Jan '04

Following on from the comment bt Shane Kerr > Also, what various contact information means is not clear. What > administration does the "admin-c:" do exactly? Only update the RIPE > Database? Coordinate with the RIPE NCC? Is it the CEO of a company? > The manager of the IT department? Or an independent contractor who handles > all the details of connectivity? I have to agree that "what is the responsibility" or perhaps just "the role" of the admin-c and tech-c. From what I have read on RIPE documents the role is not at all clear. I believe it said the admin contact has to be at the same address as the network? So what of unattended sites? There are many questions, and perhaps there should be some attempt to define more accurately what admin-c and tech-c are meant to be used for. As internet usage is becoming more of an issue (spam, child porn, etc) I can't help feeling that there should be some formal "responsibility" associated with an inetnum (like we have a "keeper of a vehicle" in the UK who has some responsibilities). I suspect to go this far would request some legal framework in countries affected, but RIPE could at least indicate, say, "The admin contact is intended to take responsibility for the IP addresses" or some such, which would be a good basis. Of course, if there was an abuse-c, then this could be a document explaining the detailed role of admin-c, tech-c, and abuse-c. Also, the "trouble" field is a good thought - maybe we should add "trouble" to inetnum, as there is already a field name defined for the purpose essentially. Just an idea. -- Rev Adrian Kennard Andrews & Arnold Ltd / AAISP www.aaisp.net.uk

1 0

E-mail Client Testing for S/MIME Compliance
by Denis Walker 09 Jan '04

09 Jan '04

[apologies for duplicate messages] Dear Colleagues, At the RIPE 45 Meeting a proposal was presented to implement an additional form of strong authentication for protecting objects in the RIPE Database and to allow for secure communication by e-mail with the RIPE NCC. This implementation uses the X.509 certificates. A question was raised by some members about how widespread is the compliance of S/MIME within mail clients. The RIPE NCC Software Engineering Department has undertaken a study to evaluate the use of S/MIME by mail clients. The results of this study can be found on our web site at: http://www.ripe.net/ripencc/pub-services/db/mail_client_tests.html Regards, -- Denis Walker Software Engineering Department RIPE NCC

4 5

Domail Clean-up
by RIPE Database Manager 06 Jan '04

06 Jan '04

Dear Colleagues, As of 6 January, 2004, messages concerning inconsistencies between "nserver:" attributes in DOMAIN objects and delegation NS RRs in our zone files will be sent out. The messages will be mailed to various contact addresses for the relevant DOMAIN objects. In addition to the description of the inconsistencies, an indication of the 'default actions' will also be provided. We kindly request data be updated so that the inconsistencies are corrected before 1 March, 2004. About two weeks before 1 March, 2004 we will send a second set of warnings for the remaining inconsistencies. By 1 March, 2004 all remaining inconsistencies will be cleaned by the RIPE NCC, using the 'default actions'. If you receive a warning about inconsistencies we recommend that you fix the data yourself. It may be that you receive multiple mails for multiple domains with multiple inconsistencies. If you want to configure a dedicated mail filter the messages will contain the subject header "Subject: Reverse DNS inconsistencies in the RIPE Database" and originate "From: <ripe-dbm(a)ripe.net>". If you have further questions please contact <ripe-dbm(a)ripe.net>. Background The cleanup is part of the effort to streamline reverse DNS operations [1] and make it easier for network range users to manipulate related DOMAIN objects. The goal of the cleanup [2] is to streamline the contents of DOMAIN objects and zone configuration files so that we can create zone files from the Whois Database. An explanation of the various possible inconsistencies can be found at: http://www.ripe.net/reverse/rdns-project/Cleanup.html. ----- [1]: Original proposal http://www.ripe.net/reverse/proposal.html. Discussion on this proposal has taken place on the RIPE NCC Services WG mailing list and was summarized in: http://www.ripe.net/ripe/mail-archives/ncc-services-wg/2003/msg00361.html [2]: Original cleanup proposal http://www.ripe.net/ripe/mail-archives/db-wg/2003/msg00738.html. ----- Can Bican Software Engineering Department Olaf Kolkman New Projects Group RIPE NCC

1 0

creating a mntner
by Randy Bush 29 Dec '03

29 Dec '03

so, i have a clean database, 3.2.0 server running, blah blah blah now i have the old problem of creating a first mntner: i run bin/dbupdate -c conf/rip.config -t -d mntner: MAINT-RGNET descr: RGnet admin-c: RB366 upd-to: rw(a)rg.net auth: NONE override: <cleartext override password> mnt-by: MAINT-RGNET referral-by: MAINT-RGNET changed: randy(a)psg.com 20031217 source: RGNET ^D and it pukes saying Create FAILED: [mntner] MAINT-RGNET ***Info: Syntax check passed mntner: MAINT-RGNET descr: RGnet admin-c: RB366 upd-to: rw(a)rg.net auth: NONE mnt-by: MAINT-RGNET referral-by: MAINT-RGNET changed: randy(a)psg.com 20031217 source: RGNET ***Error: Unknown object referenced and the interesting part of the debug output up_source_commit: entered with operation [UP_CREATE] up_send_object_db: entered, operation [UP_CREATE] up_send_object_db: object to be sent [ mntner: MAINT-RGNET descr: RGnet admin-c: RB366 upd-to: rw(a)rg.net auth: NONE mnt-by: MAINT-RGNET referral-by: MAINT-RGNET changed: randy(a)psg.com 20031217 source: RGNET ] up_get_transaction_id: entered spent 0.00 sec; got 1 rows from [dbrgnet: INSERT INTO tid VALUES(NULL)] up_get_transaction_id: exiting with new id [18] up_send_object_db: transaction ID [18] SK_write = { sockfd=[7], buf=[ADD], count=[3] SK_write = { sockfd=[7], buf=[ ], count=[1] SK_write = { sockfd=[7], buf=[18], count=[2] SK_write = { sockfd=[7], buf=[ ], count=[2] SK_write = { sockfd=[7], buf=[mntner: MAINT-RGNET descr: RGnet admin-c: RB366 upd-to: rw(a)rg.net auth: NONE mnt-by: MAINT-RGNET referral-by: MAINT-RGNET changed: randy(a)psg.com 20031217 source: RGNET ], count=[218] SK_write = { sockfd=[7], buf=[ ], count=[2] SK_write = { sockfd=[7], buf=[ACK ], count=[4] SK_write = { sockfd=[7], buf=[18], count=[2] SK_write = { sockfd=[7], buf=[ ], count=[2] up_send_object_db: send complete up_send_object_db: received from socket [ %ERROR 4 ] up_send_object_db: received from socket [ E[4][1:RB366]:ref integrity: reference cannot be resolved ] up_send_object_db: received from socket [ ] up_send_object_db: received blank line from socket, stop reading Closing socket... 7 up_send_object_db: exiting with value UP_OK up_interpret_ripudb_result: entered, ripupd_result [%ERROR 4 E[4][1:RB366]:ref integrity: reference cannot be resolved clue by four, please randy

4 11

Re: [db-wg] 3.2.0 make_db getvar cores
by Randy Bush 19 Dec '03

19 Dec '03

> Everything seems OK in your config files you've sent. I could not > replicate your problem. My best bet is that, in rip.config > SOURCEFILE /usr/home/whois/conf/sources.config > points to a file which is _not_ the one you've sent. sigh. with that gdb output? puhleeze! (gdb) back #0 0x805087d in ca_parseSrcLine (lineStr=0xbfbff74c "database:\tdbrgnet", srcStructPtr=0x80c8400) at ca_sourceLoader.c:792 #1 0x8050147 in ca_readSources ( sourcesDefFile=0x80c2840 "/usr/home/whois/conf/sources.config", confVars=0x80b1840) at ca_sourceLoader.c:464 #2 0x804faa2 in ca_init (configFile=0xbfbff9b1 "../conf/rip.config") at ca_initFn.c:87 #3 0x804bc2f in main (argc=7, argv=0xbfbff858) at getvar.c:84 let me try to debug by turning on debugging in ca_sourceLoader.c # /usr/home/whois/utils/getvar -s SAMPLE -c /usr/home/whois/conf/rip.config -p dbhost Inside ca_readSources() function. Sources file is: /usr/home/whois/conf/sources.config We are reading a comment or an empty line ..... line:DATABASE SAMPLEDB:End of line Length of line: 17 LINE >>>DATABASE SAMPLEDBEND_OF_LINE Beginning of a database defintion ..... Database name is: SAMPLEDB line:host: localhost:End of line Length of line: 16 LINE >>>host: localhostEND_OF_LINE tokens[0] = host tokens[1] = localhost VarName: host; dbComp: localhost line:port: 3306:End of line Length of line: 11 LINE >>>port: 3306END_OF_LINE tokens[0] = port tokens[1] = 3306 VarName: port; dbComp: 3306 line:trx_support: n:End of line Length of line: 14 LINE >>>trx_support: nEND_OF_LINE tokens[0] = trx_support tokens[1] = n VarName: trx_support; dbComp: n line:user: rip:End of line Length of line: 10 LINE >>>user: ripEND_OF_LINE tokens[0] = user tokens[1] = rip VarName: user; dbComp: rip line:password: secret:End of line Length of line: 19 LINE >>>password: secretEND_OF_LINE tokens[0] = password tokens[1] = secret VarName: password; dbComp: secret line:/DATABASE:End of line Length of line: 9 LINE >>>/DATABASEEND_OF_LINE We have reached the end of a DATABASE definition Testing the population of the db structure: localhost::3306::n::rip::secret:: ^ wowie zowie! no database name ^ the very problem we sought guess what happens when the password is nine characters! .../ripe-dbase-3.2.0# grep -c strcpy ./modules/ca/ca_sourceLoader.c 16 oh wonderful and thanks a million for the debugging help randy

1 0