- db-wg - mailman.ripe.net

Re: ** URGENT ** New database software at the NCC
by Wilfried Woeber, UniVie/ACOnet 08 Sep '93

08 Sep '93

Marten, >Just some explanation, the messages sent to assign are first checked to see >if they are already in the database, and currently bounced back to the NCC if >they are. I think what I will do is create something like auto-assign, which >has the same "on-the-fly" features as auto-dbm, but will only ADD new >objects, and bounce all objects that are already present. This would mean >that also contact people that are already present will not be updated when >sent to auto-assign. ...does this mean that we are no longer able to submit *updates* for person objects along with a new network object from the 193,194 address range? If this is the case, then I think we should generally go for a system where *all* objects *with the exception of objects for newly assigned networks* should go to the ripe-dbm mailbox. Wilfried.

1 0

** URGENT ** New database software at the NCC
by Marten Terpstra 08 Sep '93

08 Sep '93

Dear people, You have all probably heard that new RIPE database software has been written at the NCC. We are convinced that this software should be used as soon as possible at the NCC, because the current update procedure has had several changes made already to keep it running with the current size of both the database and the amount of updates we receive. Therefore, we have planned to start using the new software at the NCC this Friday September 10th, at 10 AM. The new software has some consequences for whois users and people that send in updates. These consequences have been listed below. Please read the document below carefully. If you have any questions or major objections, please contact me, Daniel or Tony. Cheers, Marten PS All updates sent to the new mail alias before Friday 10 AM will either bounce, or update some test database. These updates will NOT be incorporated in the real RIPE database. ----- Introduction The software for the RIPE database has recently been completely rewritten, and the new implementation will have some effects on the behaviour of the database software as compared to the software currently in use at the NCC. This document describes the impact the new RIPE database software will have on usage of the database and on updates to it. USER IMPACT With the new database software, there will be some changes in the bahaviour of the database whois server. These include: New -s flag The indexed database will still contain information from various sources. As before normally only the RIPE information will be shown. The flag for requesting information from all sources remains the same (-a), and a new flag will be introduced to specify one or more specific source (-s source). The RIPE whois client available from the RIPE document store has been modified to accept this flag. Recursive Lookups Limited to Source of Reference The "-a" flag, or asking for information from multiple sources will show a slightly different behaviour. The old software used to perform recursive lookups on ALL the selected sources. This means that a contact person referenced by a NIC entry would be looked ap and listed from ALL other sources. The new software will ONLY perform recursive lookups in the source where the reference was found. Consequence is that one does not get a mix and match of referenced information from various sources. All references are ONLY looked up in the source where they are referenced. This was done to limit output volume and user confusion. Strict Enforcement of Object Definitions The current database file has objects in it that do not conform to the official object definitions. The new whois server will only output the legal fields for such an object. So, it may be that a network object contains a zone-c attribute in the database file, but the whois server will never show this attribute, because it is not in the network object definition. An overview of the defined attributes per object can be found at the end of this document. UPDATE IMPACT There are a number of changes that will affect people sending in updates for the RIPE database. Real Time Updates The new database software works with on-the-fly updates. This means that accepted and acknowledged database updates will be visible immediately in the whois server running at the NCC. These changes will not be visible immediately in the FTPable database file maintained via the RIPE Document Store. Once a day -or night rather- an automatic "cleanup" procedure will run which generates the FTPable file and adds all the guarded attributes as specified in the guardian files maintained at the NCC. This means that the database available for anonymous FTP does not necessarily have the same information as the running database. Guarded attributes will not be changed using the normal update procedure, they will be added once a day, and remain unchanged until the next cleanup time. New Update Procedure A new mailbox will be installed for direct access to the new database software without RIPE NCC staff intervention. This means fast response times, but since the updates are not looked at by the staff, updates containing errors will be bounced just as fast. The new mailbox will be <auto-dbm(a)ripe.net>. The old mailboxes will remain, and the NCC staff will look at mails directed to one of these mailboxes and forward updates to the new update procedure, after some very basic checking. The NCC will NOT correct all syntax errors in updates sent to one of these mailboxes, but it might correct some obvious errors. As before all updates and acknowledgements will be logged at the NCC. This means that errors and "strange" updates can be tracked quite easily. Stronger Syntax Checking The syntax checking in the new software has been improved and strengthened. As with the old software, database objects that cause errors will be bounced, and NOT be incorporated in the database. Warning messages are usually used to indicate corrections that have been made to the object sent in. Objects with warnings WILL be incorporated in the database, with the changes made as indicated in the warning messages. In general error and warning messages are descriptive and will give some indication where an error or warning was generated. The new syntax checking has the following rules: Database objects that miss mandatory attributes will generate an error, and will be bounced. An overview of the defined attributes per object can be found at the end of this document. Objects that have attributes that are not defined in the object, will generate an error and will be bounced. Objects that have syntax errors in the values of the various attributes will generate an error, and will be bounced. Objects that have a date in the changed field that is prior to the date in the current database entry will generate an error and will be bounced. Guarded attributes in updates will be ignored other than generating a warning if the values do not match the current value in the database. Currently only the boundary gateway and routing privilege in the network object are guarded attributes. The aut-sys attribute in the network object will be turned into a guarded attribute as soon as possible. Deletions have been made more easy to perform. To delete an object, it is sufficient to send in that object plus a delete: attribute that describes why this object has to be removed from the database. Objects that need to be deleted, will NOT be checked for syntax, but must exactly match the object currently in the database, otherwise an error will be generated, and the object will not be removed from the database. The order of the different attributes in the updates is NO LONGER relevant for the match. The order of different lines in the same attribute is. The new syntax rules have been determined from the original object definitions. It may well be that some syntax rules have been made too strict, or are simply wrong. Please do not hesitate to contact us in these cases. The syntax is not fixed forever. Attribute definition per object as used by the new database software. inetnum object Mandatory: inetnum netname descr country admin-c tech-c connect changed source Optional: aut-sys (will be guarded asap) gateway (obsolete) nsf-in nsf-out (obsolete ?) routpr-l (guarded) bdrygw-l (guarded) remarks rev-srv person object Mandatory: person address phone changed source Optional: e-mail fax-no nic-hdl remarks domain object Mandatory: domain descr admin-c tech-c zone-c changed source Optional: dom-net nserver sub-dom remarks aut-num object Mandatory: aut-num descr admin-c tech-c changed source Optional: guardian as-in as-out default remarks bdry-gw and rout-pr will be obsolete in the near future.

2 2

misleading use of the connect: field in inetnum: object
by Wilfried Woeber, UniVie/ACOnet 07 Sep '93

07 Sep '93

Dear all, (and sorry for a possibly too wide distribution). We had some instances recently, where operations people were prompted by users to sort out connectivity problems to a given network that couldn't be reached. The most obvious thing to do is to look into the RIPE-DB, of course. The network(s) were registered, but further investigation revealed that the network(s) did not yet have full connectivity to the Internet. However the information in the connect: field of the network object contained strings like RIPE, NSF or EASI, but no indication about the more local situation. For the end-user, and most probably for a lot of operations people as well, it is definitely not obvious that the connect field is more or less just a (sometimes very misleading!) comment. Thus I propose to discuss and agree on a (RIPE) recommendation along the following lines, at least for registering new networks in the database: - The Local-IR registers the network with a (local/regional) network or service-provider string, but without using any misleading flags in the connect: field like RIPE, NSF, EASI... If -for whatever reason- commonly used strings are used, then the remarks: field should give a very clear indication about expected connection date(s) and service restrictions (like dial-up networks, e-mail only links...). - When the network is connected to a service provider to configure routing, (Reverse-)DNS, or services like e-mail, etc. the network should be tagged with a useful service-provider string. This "state" could and probably should be skipped for situations where full connectivity could be expected within a few days. Examples: ACONET or SURFNET for national academic networks, EUNET-DE for a EUnet connection, XLINK and the like Note: I'm not proposing to use the LOCAL tag, because this means that no connectivity to the outside world is wanted or intended at all! - Only after the point in time where there is really full connectivity, the usual strings like RIPE, NSF, etc. should be added to the connect: field, out-dated comments removed, DNS and AS-Number information added, and the like. In addition to that I propose to list the connect: strings in a kind of intuitive "local to global" sequence. Examples: "ACONET RIPE NSF" or "ARNES EMPB RIPE NSF" or "EUNET-XX RIPE" Thus we would end up with really useful information in the connect field, and maybe also with more complete database objects, because information is available for the second or last update phase that usually cannot be submitted when registering the network. Could I have your opinions and comments please? An example for a network in Western-Nowhere, that is connected to a regional net that gets connectivity throug ACOnet and EBONE would be: Phase-1 connect: WNWNET !while setting up the net !or eg. e-mail only conn Phase-2 connect: WNWNET ACONET RIPE !after adding... bdrygw-l: ACONET Phase-3 connect: WNWNET ACONET RIPE NSF !after registering NSF-conn. Wilfried. ------------------------------------------------------------------------------- Wilfried Woeber : Wilfried.Woeber(a)CC.UniVie.ac.at (Internet) Computer Center - ACOnet : Z00WWR01(a)AWIUNI11.BITNET (EARN) Vienna University : 29133::WOEBER (SPAN/HEPnet) Universitaetsstrasse 7 : Tel: +43 1 436111 355 A-1010 Vienna, Austria, Europe : Fax: +43 1 436111 170 NIC: WW144 -------------------------------------------------------------------------------

3 2

doc update for the (new) RIPE DB-Software
by Wilfried Woeber, UniVie/ACOnet 01 Sep '93

01 Sep '93

Dear DB-interested folks, according to action 15.11 on me from the 15th RIPE meeting, and as the beta test phase for the new DB-Software is making good progress, we have to agree on the structure, contents and production of new or updated documentation as well. Thus I ask all of you to contribute to this discussion on the DB-WG list. A copy of this is also sent to the Local-IR list FYI, however I propose to have the real discussion going on on th DB-WG list. Amongst other things, I think we should agree on - what is the intended audience for the doc? - what is a good structure? - who should contribute and how/when do we produce the paper(s) Regards, Wilfried. ------------------------------------------------------------------------------- Wilfried Woeber : Wilfried.Woeber(a)CC.UniVie.ac.at (Internet) Computer Center - ACOnet : Z00WWR01(a)AWIUNI11.BITNET (EARN) Vienna University : 29133::WOEBER (SPAN/HEPnet) Universitaetsstrasse 7 : Tel: +43 1 436111 355 A-1010 Vienna, Austria, Europe : Fax: +43 1 436111 170 NIC: WW144 -------------------------------------------------------------------------------

2 1

Authorisation and Notification of Changes
by Daniel Karrenberg 01 Sep '93

01 Sep '93

Folks, here is my long promised writeup on the "notify" and "maintainer" attributes. I hope it is clear enough. It will be discussed at the next RIPE meeting. Hopefully no substantive changes will be necessary. Suggestions on how to improve clarity are always welcome of course. Daniel Authorisation and Notification of Changes in the RIPE Database Daniel Karrenberg RIPE NCC ABSTRACT Two new attributes for all objects in the RIPE database are proposed in order to implement a generalised method for authorising changes and to notify interested parties of any changes made to a specific object. In addition the authorisation method provides a convenient way for distributed maintenance of the database. The Notify Attribute Each RIPE database object will have an optional attribute called notify. The value of the notify attribute is one valid RFC822 e-mail address. There can be multiple notify attributes. Whenever the object concerned is changed in the database a notification message will be sent to each e-mail addresses appearing in a notify attribute. This makes it straightforward to keep track of changes to specific objects and prevent changes from going unnoticed. Multiple notify attributes make it possible to notify a number of interested parties. This could be used to alert all contact persons for an object or the local contact per- sons as well as the relevant service provider. Although it may be tempting to put many notify attributes on database objects in order to notify everyone even remotely interested, this is not recommended. A very few key addresses should be sufficient. Prior to entering any mail address here, the explicit or implicit consent of the person responsible for that particular mailbox needs to be obtained. - 2 - The Maintainer Attribute Each RIPE database object will have an optional attribute called maintainer. The value of the maintainer attribute is a registered maintainer name. There can only be one main- tainer attribute per object. Whenever a change to the object concerned is attempted in a copy of the database the maintainer attribute of the current database object is examined. If there is no maintainer attribute or the maintainer name is authorised to make changes in the copy of the database the update proceeds causing the necessary notifications as per the notify attribute. If the maintainer name has no authorisation to change the local copy of the database, the update request is forwarded to the maintainer for processing. No notifications are per- formed in this case. The following data needs to be maintained locally about each maintainer: Maintainer name Authority none change whole database change only own objects Forwarding Info mail/RFC822-address other/address Authorisation none mail/RFC-822-address other/key Application 1: Regional Registries In order to align the InterNIC and RIPE databases it has been agreed that European objects will be maintained in Europe. The RIPE NCC will provide the data for these objects to the InterNIC for inclusion in their database without further processing. The RIPE NCC will refer all updates for non-European objects to the InternNIC and the InterNIC will refer all updates for European objects to the RIPE NCC for processing. This can be achieved by creating two maintainer names: - 3 - INTERNIC and RIPE-NCC and tagging all European objects with RIPE-NCC and vice versa. The tags can be phased in slowly, avoiding a flag day with the associated massive consistency problems. Over time all objects in the RIPE database would be thus tagged. Updates from third parties for objects with the maintainer attribute added can now be referred correctly. Updates from the other registry for objects it maintains can be accepted without further checking. Example 2: Local Registries Some European local registries keep their own copies of the database containing the objects within their area. This leads to consistency problems as updates can be sent both to the RIPE NCC and to the local registry. Referrals are per- formed by ad hoc methods. Frequently only one of the data- bases is updated and alignment needs to be done manually. By registering maintainer names for the local registries and tagging the appropriate objects this can be automated and made more reliable. The NCC would forward update requests for locally maintained objects to the local registry unless they come from that local registry itself. Example 3: Guarded Objects Some objects such as the autonomous-system object (see ripe-81) need to be protected against changes by anyone but a designated guardian since changes to these objects have a direct operational impact. By registering appropriate maintainer names for the guardi- ans and tagging the objects to be protected this functional- ity can be provided in a canonical way. Any change by third parties to such an object will not only be prevented but cause automatic notification of the guardian through the forwarding mechanism.

2 1

beta release new database software
by Marten Terpstra 20 Aug '93

20 Aug '93

People, I have made a distribution of the new software. It includes most of the stuff you would need for indexing and updating a RIPE like database. For those of you who have been playing with the 2 early releases, here are some of the changes: - minor change to key generation routine (to allow certain chars) - dbupdate can now be used as an email responder - more stuff made configurable, more configuration options - some general bug fixes - a bit more documentation (I know, still nowhere near perfect) You will need some familiarity with the RIPE database structure to get this stuff running without much hastle. I have tried making it into a read, install and run package, but it is not quite really. Please understand that it IS still beta, and there may well be bugs everywhere, and it may not be as configurable as you would like, but I'm working on that. I just wanted to get this thing out so that you could play with it a bit, and see if you like it. If you do not have some time to spare to install and try to understand it, I would advide not to get the stuff just yet, but wait for the first real release (expected some time around the RIPE meeting). This release does not understand guarded fields yet, all hooks are in, just some code missing. The indexer and whoisd have been tested by some people already and did not show any major bugs. The update mechanism is not tested too well yet, so try building some test databases first to see how that all works. Most of the working and installation is described in 2 files, INSTALL and README. README should explain enough to get you started, experiment after that. You can find the stuff on: ncc.ripe.net:dbase-beta/dbase-beta.tar.Z Directory dbase-beta is unreadable, but you CAN fetch the file and note that this is NOT on ftp.ripe.net. Oh, you need PERL to run all this stuff ... Enjoy, -Marten

1 0

RIPE Database beta-software trials ...
by Marten Terpstra 16 Aug '93

16 Aug '93

[Apologies for all who received this message on the dns-wg list already, that was obviously a mistake ... -Marten] Dear all, We finally have some beta software running at the NCC for the database, and we would like you to test some things. We have created an empty test RIPE database on ncc.ripe.net running with the new whois daemon. You can send in updates to: testdb(a)ripe.net It will be processed and incorporated in a test database running on ncc.ripe.net (also a whois daemon running on that database). You should receive a ack mail back (to the Reply-To field, or From: is no Reply-To field present) and you should be able to find it in the whois server immediately. Acks should be sent to you within minutes. You can also delete objects in that test database, by simply adding a delete: line to an object. So, person: Marten Terpstra address: RIPE Network Coordination Centre (NCC) address: Kruislaan 409 address: NL-1098 SJ Amsterdam address: Netherlands phone: +31 20 592 5065 fax-no: +31 20 592 5090 e-mail: marten(a)ripe.net nic-hdl: MT2 changed: marten(a)ripe.net 930503 source: RIPE delete: just try and delete this thing should delete this object in the database, if all fields are the same as in the database. The ordering of fields does not matter, just as long as attributes with the same name are in the same order, ie in the object above you cannot change the place of two "address:" fields. Also the place of the delete field does not matter, it can be anywhere in the object. Updating should be very quick, ack mails should be sent to you within minutes. Please try some stuff and see what happens. Comments on the format of the ack mail are also appreciated. Currently verbose ack is default ie you get at least a one line ack per object you send in, and the full object if there were errors or warnings. If you send in objects with a source other than RIPE it should complain that this is an unknown source and should refuse the update. -Marten PS All requests are already properly logged so be careful with those deletes ;-) PPS The syntax checker for this test database is VERY strict, simply because there is no manual intervention from the NCC any longer. It may well be that existing objects in the real database are no longer accepted ;-)

1 0

RIPE database software beta sites needed
by RIPE Database Management 04 Aug '93

04 Aug '93

Folks, Finally parts of the new RIPE database software are ready for some beta testing. I would like a not too big number of sites/people that are willing to run some tests on the new software, and actually have some time to spend on this to make themselves known to me. I have created a new (temporary) mailing list <db-beta(a)ripe.net> which will deal with all the beta-testing on the software. That is where I will put details on where to get the software and will send bug fixes. If you are willing to participate in the beta testing, please send a mail to db-beta-request(a)ripe.net or simply reply to this message. If you just want to see what is going on with the beta testing, you are welcome to join the list as well. Just hope that not all of you will be just listening in ;-) Suggestions/comments welcome. -Marten PS Some people already indicated they wanted to join the list, so currently I have the following people on the db-beta list: Wilfried Woeber <Wilfried.Woeber(a)CC.UniVie.ac.at> # db-wg chairman Tony Bates <tony> Marten Terpstra <marten> GARR NIS <info(a)nis.garr.it>

1 0

Re: The addition of guarded fields
by farrache＠ccpnxt5.in2p3.fr 27 Jul '93

27 Jul '93

Marten, I think that the RIPE NCC have to split blocks. RIPE NCC is a neutral organisation that can do it without problems. If we do not agree on that, that means a very long procedure to update the guarded fields, and I believe that no service provider will want to rely on the efficiency of a NIC for its routing policy. So that mean that nobody will use those fields for routing, so the information may be completly out of date. But I agree with Blasco on one point, the maintainer of the block must be inform that the block has been splitted Gilles

3 2

Re: Unique NIC/RIPE/xxx-Handles
by scottw＠internic.net 27 Jul '93

27 Jul '93

If a centalized number is important we have completed and are testing software that will allow other NICs to get InterNIC handles for their users. Access to the InterNIC handle generator will be controled by the source address of the host requesting the handle. I strongly beleive that handle generation below the regional registry would be dangerous. To resurve a handle we only require the name and email address. Scott

1 0