From nitrox at via.ecp.fr Tue May 1 00:27:21 2001 From: nitrox at via.ecp.fr (Damien Lucas) Date: Tue, 1 May 2001 00:27:21 +0200 Subject: [ANN] VLANserver 0.0.1 Release Message-ID: <20010501002721.A11064@via.ecp.fr> The VLAN server is now released. This is the server which enables you to change the channel of a client. This is a workaround for networks which are not multicast capable. For more questions about what this server is doing please read the FAQ first. We have to make you aware that this server is NOT secure, that means that we are not responsible if you get troubles with it. Because of this security problems please do not make any binary or packages ! If you want to debug any help is welcome but be aware that we are also developping another one from scratch which will have more features and be more secure. You can dowload the source of the VLAN server in a tar.gz via HTTP or FTP at : http://www.videolan.org/pub/vlcs/0.0.1/VLANserver.tar.gz ftp://ftp.videolan.org/pub/videolan/vlcs/0.0.1/VLANserver.tar.gz If you have any question mail us at : vlcs at videolan.org Have a nice day. For the VideoLAN team. -- nitrox From cvs at videolan.org Tue May 1 01:32:10 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Tue, 1 May 2001 01:32:10 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010430233210.CF9282340B@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: .cvsignore ChangeLog TODO Update of /cvs/videolan/network/VLANserver/doc Added Files: passwd.sample Update of /cvs/videolan/network/VLANserver/interface Modified Files: telnet.c Log Message: . ChageLog and TODO list reformatted . Corrected a bug in authentification -- nitrox From nitrox at via.ecp.fr Fri May 4 08:58:20 2001 From: nitrox at via.ecp.fr (Damien Lucas) Date: Fri, 4 May 2001 08:58:20 +0200 Subject: Guidelines for new vlcs Message-ID: <20010504085820.A17115@via.ecp.fr> A short summary of the decisions we took yesterday about the architecture of the new vlcs: NETWORK: -------- NET_init() NET_destroy() NET_addto_channel() NET_removefrom_channel() NET_user_info() NET_machine_realinfo() NET_update() INTF: ----- provides: calls: IF_init() IF_destroy() IF_reload_module() CS_reload_module() IF_reload_submodule() IF_SM_start(), IF_SM_stop() IF_addto_channel() CS_addto_channel() IF_removefrom_channel() CS_removefrom_channel() IF_get_user_info() CS_get_user_info() IF_get_machine_realinfo() CS_get_machine_realinfo() IF_update_db() CS_update_db() IF_authentificate() CS_authentificate() CONFIG: ------- CFG_init() CFG_start() CFG_stop() CFG_destroy() CFG_dump() DATABASE: Here is more a list of features than a list of functions --------- Creation/Destruction of a table Add/delete/search an entry Look up of the whole database Transactions not multi-keys locks with priorities (by time) save / load table Internal representation of data is hidden (only the size is known by DB) Here are some more problems we found out: - How to identify a computer ? - How to identify a user ? - Database update ? - How to give config parameters to submodules - Should we have a submodules' list written somewhere ? - Which method to use for reload of sub-modules ? - init/destroy/start/stop - user_info: db/net ? - sub-modules "interface" ? - Where and who we decide the user's rigth's for programs ? - How to manage channels' methods (vls)? OK. Just say if I wasn't clear enough ! Next time is Thursday 10th May 21h -- 'trox From henri at via.ecp.fr Fri May 4 10:22:05 2001 From: henri at via.ecp.fr (Henri Fallon) Date: Fri, 4 May 2001 10:22:05 +0200 Subject: Guidelines for new vlcs In-Reply-To: <20010504085820.A17115@via.ecp.fr>; from nitrox@via.ecp.fr on Fri, May 04, 2001 at 08:58:20AM +0200 References: <20010504085820.A17115@via.ecp.fr> Message-ID: <20010504102205.D14804@via.ecp.fr> On Fri, May 04, 2001, Damien Lucas wrote: > - Database update ? Is it possible to have it segmented, let's say for example one segment for one switch, and have a lock by segment. When you want to update, you lock the first segment, request related to that segment stay in a queue, snmp renew the informations related to this segment, then you delock it and process the request ? Two advantages : - request are queued and then processed, not lost - only a very small part of the db in not available at one time, and having it segmented makes it more flexible. Is this idea a bad idea ? -- henri From bbp at via.ecp.fr Fri May 4 10:56:56 2001 From: bbp at via.ecp.fr (Brieuc Jeunhomme) Date: Fri, 4 May 2001 10:56:56 +0200 Subject: Guidelines for new vlcs In-Reply-To: <20010504102205.D14804@via.ecp.fr>; from henri@via.ecp.fr on Fri, May 04, 2001 at 10:22:05AM +0200 References: <20010504085820.A17115@via.ecp.fr> <20010504102205.D14804@via.ecp.fr> Message-ID: <20010504105656.A425@via.ecp.fr> > Is this idea a bad idea ? This is not a bad idea, but it has something to deal with the problems mentionned by nitrox, we have not decided yet how we will identify a set of machines/switches/... so maybe we will be able to do this, but this is not sure yet. The question has been raised during the discussion. Anyway, even if it would be nice, it wion't be required, because the database will not be "down" during its update, for instance, if you have a switches table, the update will take one switch after another, issue snmp requests to collect data about the switch, then lock the switch, write the data, and unlock the switch, then take the next swqitch... There won't be any downtime during database update. -- BBP From cvs at videolan.org Wed May 16 00:57:23 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Wed, 16 May 2001 00:57:23 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010515225723.93EC72340F@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: logger.c logger.h types.c Update of /cvs/videolan/network/VLANserver/config Modified Files: config.fl Update of /cvs/videolan/network/VLANserver/doc Modified Files: vlanserver.conf.sample Log Message: * Make the logs wrap to 80 characters * Change a bit the way the 'method' variable is used, it is now a bit-mask to know which are the outputs : LOG_FILE or LOG_SCREEN are the two options implemented * the bit-mask change is completely transparent with old config file, the values are the same, only the method to determine is changed -- marcari From cvs at videolan.org Wed May 16 12:48:32 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Wed, 16 May 2001 12:48:32 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010516104832.683CF2340F@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: TODO Log Message: forgot to put my job as done in the todo list -- marcari From cvs at videolan.org Mon May 21 01:12:05 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Mon, 21 May 2001 01:12:05 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010520231205.8513A2340F@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: AUTHORS ChangeLog TODO logger.c logger.h types.c Update of /cvs/videolan/network/VLANserver/config Modified Files: config.fl config.h Update of /cvs/videolan/network/VLANserver/doc Modified Files: vlanserver.conf.sample Log Message: Done a new logging system (easily adaptable to the new vlcs) * logmethod has been removed * logs are sent to a list of outputs (with a type for each) * only one logfile can be used because of the way config is managed now * log outputs are referenced by their FILE* (for removing for eg) * it allows telnetd to draw logs on its ouput using FIFOs -- marcari From cvs at videolan.org Fri May 25 12:08:52 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Fri, 25 May 2001 12:08:52 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010525100852.E4D692340F@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver/server Modified Files: send.c Log Message: Humm! Just a word, before committing anything, just 1- see if it compiles 2- check with cvs diff what you change in which files 3- commit all the files where you did significant changes That shouldn't happen any more: server/../logger.h:46: parse error before `FILE' -- nitrox From cvs at videolan.org Mon May 28 02:49:41 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Mon, 28 May 2001 02:49:41 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010528004941.6AD1E23414@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: ChangeLog TODO logger.c logger.h Update of /cvs/videolan/network/VLANserver/interface Modified Files: telnet.c Log Message: Logs are now available on the telnet interface. -- marcari From cvs at videolan.org Wed May 30 01:46:03 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Wed, 30 May 2001 01:46:03 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010529234603.9AABB23413@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver Modified Files: ChangeLog TODO Update of /cvs/videolan/network/VLANserver/snmp Modified Files: callback.c if_snmp.c if_snmp.h snmp.c snmp_switch.h Log Message: . Added the close of snmp sessions after the walks . Updated ChangeLog and TODO -- nitrox From cvs at videolan.org Wed May 30 07:39:26 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Wed, 30 May 2001 07:39:26 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010530053926.DDC8823413@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver/server Modified Files: request_handler.c Log Message: . Added a test in request_handler that was the cause of a double segfault (VLANserver and VLANbridge). -- nitrox From cvs at videolan.org Wed May 30 23:53:08 2001 From: cvs at videolan.org (cvs at videolan.org) Date: Wed, 30 May 2001 23:53:08 +0200 (CEST) Subject: CVS Commit: network Message-ID: <20010530215308.4BDD223413@brahma.via.ecp.fr> Update of /cvs/videolan/network/VLANserver/interface Modified Files: interface.c Log Message: . Added a call to DB_machines_unlock after DB_machines_lock -- nitrox