![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to BUGLIST CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [CSRG BSD Unix] / 43BSDReno / old / berknet|
Return to BUGLIST CVS log | Up to [CSRG BSD Unix] / 43BSDReno / old / berknet |
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.