Return to BUGLIST CVS log

Up to [CSRG BSD Unix] / 42BSD / old / berknet

Annotation of 42BSD/old/berknet/BUGLIST, revision 1.1

1.1     ! root        1: Network Status                 November 5, 1980
        !             2: 
        !             3: BUGS
        !             4: ----
        !             5: --     Various response messages are lost.  This includes "fetching"
        !             6:        files when the file being retrieved never arrives.  I suspect
        !             7:        this has something to do with unreliable delivery of error messages,
        !             8:        but this is not reliably reproducible.
        !             9: 
        !            10: --     The net command will create files in the queue directories
        !            11:        without the corresponding control files ("dfa..." without "cfa...").
        !            12:        Unknown cause.  They should be periodically removed.
        !            13:        (Perhaps caused by an error such as an invalid machine name.)
        !            14: 
        !            15: --     The network makes no provision for errors in transit on intermediate
        !            16:        machines, such as "No more processes" or "File System Overflow".
        !            17:        While these occur only rarely, when they do, no message or
        !            18:        notification is sent to anyone.
        !            19: 
        !            20: --     The network daemons occasionally core dump.  They should not.
        !            21:        
        !            22: 
        !            23: SUGGESTIONS
        !            24: -----------
        !            25: 
        !            26: --     Maintenance Improvements:
        !            27:        The network has become large enough to make re-compilation
        !            28:        of the source on all machines to become practically impossible.
        !            29:        The net command has compiled within it a routing table for each
        !            30:        remote machine (defined in config.h).
        !            31:        Adding a new machine to the network requires recompiling the
        !            32:        net command on ALL machines.  The net command should read an
        !            33:        external text file to compute its data structures.
        !            34:        There is a program patchd, written by Bill Joy, which could
        !            35:        be used to patch the binary versions of the network
        !            36:        on like-systems, such as the Computer Center machines.
        !            37:        The network code should use the retrofit library for
        !            38:        non-Version 7 systems.
        !            39: 
        !            40: --     The possibility of a number of small UNIX personal machines wanting
        !            41:        intermittent access to the network looms ahead.  We should attempt
        !            42:        to organize the software to allow occasional use
        !            43:        by other UNIX machines, without tying down a port all the time.
        !            44: 
        !            45: --     Bob Fabry has suggested the "machine" be generalized to imply a 
        !            46:        machine/account pair, e.g. -m caf would imply "caf" on Cory,
        !            47:        -m Cory would imply "fabry" on Cory.
        !            48:        Environments could provide this information.   
        !            49:        It has also been suggested that the notion of a "default" machine
        !            50:        is too restrictive and that each type of command should have a
        !            51:        default machine, e.g. netlpr to A, net to B, netmail to C, etc.
        !            52: 
        !            53: --     Colin has developed some data compression algorithms.  On machines
        !            54:        which are normally CPU idle, his algorithms could be used to
        !            55:        compress data and speed up file transfer.
        !            56:        Each individual host could decide whether data should be compressed,
        !            57:        and each receiving machine would be able to handle both compressed
        !            58:        and uncompressed data.
        !            59: 
        !            60: --     Files being retrieved, or fetched, are created zero-length
        !            61:        as the request is sent to the remote machine.  An alternative 
        !            62:        would be to put the message "File being transferred." in the file to
        !            63:        make things clearer.
        !            64: 
        !            65: --     File modes should be preserved across the network.  Currently
        !            66:        they are set to 0600 most of the time.
        !            67: 
        !            68: --     It would be nice if the rcs facilities and commands on various
        !            69:        UNIX machines with rcs links were more accessible from machines
        !            70:        without an rcs link.
        !            71: 
        !            72: --     The network was not expected to become as large as it has.
        !            73:        Not much thought was given to large networks.
        !            74:        The netq command only lists queues on the local machine,
        !            75:        but many times the user is waiting for long queues on intermediate
        !            76:        machines.
        !            77:        Likewise, once the request is forwarded to the nearest machine,
        !            78:        the netrm command will not let the originator remove the queue file.
        !            79:        Finally, a network status command telling people what the network
        !            80:        was doing would be very helpful.
        !            81: 
        !            82: --     The underlying protocol is wasteful and/or confusing:
        !            83:        * Compute a full checksum on the entire file in addition
        !            84:          to the checksum per packet now provided.
        !            85:        It is unlikely these will be changed since all the daemons
        !            86:        on the network machines would have to be changed at once.
        !            87: 
        !            88: --     The netcp command should allow the user to default one of
        !            89:        the filenames to a directory, ala the cp command.
        !            90: 
        !            91: --     File transfers, like remote mail, should be possible from
        !            92:        the Berkeley Network to the Arpanet and the Bell Research Net.
        !            93:        This is not difficult technically, but requires UNIX-like
        !            94:        stream interfaces to be written for the gateways.
        !            95: 
        !            96: --     Currently the network files being transferred are
        !            97:        copied into /usr/spool... it would be nice for
        !            98:        large files to simply use a pointer to them.
        !            99:        (To save time and space).
        !           100: 
        !           101: --     The scheduler the daemon uses is very simple.
        !           102:        It should have a way to age priorities and to "nice"
        !           103:        transfers, to be done after all normal ones are done.
        !           104:        Also, there are some network uses that are time-dependent.
        !           105:        It would be nice if certain queue files would disappear
        !           106:        at certain times, if for example, a remote machine were down,
        !           107:        given that they are no longer useful.

This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.