ITS distribution notes: Lots of random stuff in random order, some old, some new. Most of it was personal notes to myself, only vaguely intended for publication, which I never had time to clean up. As you read through this stuff, you should be thinking in the back of your head, "How can we get all this (%*@(%!# crap to work better?" Here's one suggestion. Ideally there would be a single file containing all of the runtime site-dependent config stuff (hostname, address, params for various progs, etc). Can maintain numerous configs in a single directory called SYSCNF with a link pointing from the default of "@ DEFS" to to the file actually in effect, which would be SYSCNF; . ITS would then check this default file on startup for its identity. Can also force use of a specific config file by depositing something with DDT in an otherwize zero location prior to startup (its machname for example, from which the SYSCNF file can be found). This location can even be known and set by the KLH10 init file! Notes on system-dependent config stuff: If you are generating a NEW instance of ITS with a new name, there are a lot of things you have to do. Read through this whole section! To rebuild ITS, do: :MIDAS SYSTEM;_ITS and then type the desired machine name, in UPPERCASE. There are a lot of undefined symbols in EVSYMS at the end; these can be ignored. You may want to rename the resulting ITS BIN to ITS , eg :rename its bin, its 645pi in order to avoid clobbering it with the next assembly. To get a KLH10-loadable executable from that: Extract it from ITS with either: [a] FTP it out (use image mode) to a file, eg "itsbin.645pi" [b] Virtual magtape (copy to MT0:, uses coredump mode) KLH10: devmount mta0 tapeout rw fmt=raw ITS: :copy system;its 645pi, mt0: (If get RH11 ERROR, ignore and try again) KLH10: devunmount mta0 Convert to format of choice, normally U36, with one of: FTP: ./wfconv -fu < itsbin.645pi > itsbin.645pi-u Tape: ./wfconv -cu < tapeout.tpr > itsbin.645pi-u In a fresh KLH10, no config needed: ./kn10-ks /dev/null load @.ddt-u load @.nsalv-260-u load itsbin.645pi-u dump @.its-645pi-u quit Thereafter you can just load @.its-whatever and go. NEW SYSTEM ========== Things that must be edited for each new ITS system/machine. ITS MACHINE NAME: Pick a good one. Historically all ITS machine names have been exactly two characters. Someday we may have a list of just what things actually depend on this or some other limit, as well as whether digits are permitted. @DEV - requires 2 characters for other ITSes. MLDEV - requires 2 characters for other ITSes. LISP - requires 3 characters due to CAIE checks. Code could easily be fixed for 6. Build new ITS ------------- Modify these files and then rebuild ITS per above instructions. SET NAME: The ITS machine name, two characters. SYSTEM;CONFIG > - define chunk of params for each name. SYSTEM;ITS > - add to comment at start? Not really needed. SET NET ADDRESS: SYSTEM;CONFIG > IMPUS3 - the IP address. NM%IMP - Subnet mask for IMPUS3. IMPUS - old IMP address. Set it to low-order byte of IMPUS3 to avoid breaking any old semi-working stuff.. TTY TABLE: SYSTEM;TTYTYP > Machine-name conditionalized; must have machine-specific entries for hard-wired terminal lines. Build user config files and programs ------------------------------------ HOSTNAME: Very important -- COMSAT and other things will break if they can't find their own system name in the host tables! First edit this file to add hostname(s): SYSHST;H3TEXT > Although this used to be dynamically generated, it must now be done by hand. Then run: SYSHST;H3MAKE BIN which snarfs up H3TEXT and compiles it into SYSBIN;HOSTS3 >. This can take quite a while depending on how large H3TEXT is. TELSER: SYSNET;TELSER > TELSER has hard-coded greeting messages based on hostname. If it doesn't recognize its own hostname, it uses "Unknown ITS PDP-10". This can be modified if desired (just before TCPOL:) for hack value. :MIDAS SYSBIN;_SYSNET;TELSER MLDEV: SYSEN2:MLDEV > There are two places in this program which have a hardcoded list of ITS machines. The name of the new ITS must be added here if you wish to use the MLDEV device to access the filesystems of other machines. These names must exist in HOSTS3 as it uses NETWRK"HSTLOOK to find the corresponding IP addresses. Also set DEVICE;JOBDEV symlinked to DEVICE;ATSIGN MLDEV. MLDEV/MLSLV and known job devices in general: The maximum size of a device name is 6 SIXBIT characters, although very old programs sometimes assume 3. Whenever a program uses a device name that ITS doesn't have built in, the system runs the "unknown device handler" which is SYS;ATSIGN DEVICE (source SYSENG;@DEV >). (Note the binary is a SBLK file -- it exists too briefly) It then processes this device name as follows: If "<" or ">" - Returns "No such device" error. If DSK:DEVICE;JOBDEV exists - loads and starts that program. This overrides all the rest of this stuff and is the preferable way of configuring things. Otherwise @DEV goes through a lot of trouble to figure out a plausible interpretation of the device name, using a semi-built-in table of known "foreign filesystems" and DIR device names, that it checks to map the device into one of four possible devices: DSK:DEVICE;ATSIGN DIRDEV - SYSENG;DIRDEV > DSK:DEVICE;ATSIGN MLDEV - SYSEN2;MLDEV > DSK:DEVICE;ATSIGN R.I.P. - SYSEN2;RIPDEV > DSK:DEVICE;ATSIGN RMTDEV - GZ;RMTDEV > (All are or should be PDUMPed binaries) These tables are checked as follows: If either table contains an exact match, use that. Else if the first 2 chars of the name match something in the foreign filesys table (here's one 2-char dependency!), it tries DEVICE;JOBDEV , then DEVICE;ATSIGN MLDEV. The foreign system will presumably interpret the rest of the device name (typically DIR or ARC, etc). The other table is mostly DIR where is one of ,DSK,SYS,COM,DIR,DNR. If there's still no match, it checks for a secondary pack name using a table of machine names (AI and MX for now) and a subtable of pack names for each. This seems to be only for the purpose of returning a "Pack not mounted" error if the pack is known to normally exist, as opposed to a "no such device" error. If still no match, it checks the last letter -- if it's a digit, it strips it off and starts all over again. FINALLY, it checks to see whether the request is to open the directory of a known non-directory device (ie if the FN1 and FN2 are ".FILE. (DIR)") and cobbles together a simple reply if so. The known devices for this hack are: lpt,ptr,ptp,plt,cod,dis,nvd,pdp,tab,stk MLDEV, when invoked, checks its real device name and performs some additional munching. It knows about a list of ITS systems, and for each in the list: X is transformed into . DIR is transformed into DIR (for consistency?) The 1st 2 chars (another dependency!) are then checked and if they match a , that system is contacted. list: AI ML DM MC MX MD KL KS If there's no match to any of these, then MLDEV tries one more table before giving up. If the full device name is one of the following, it substitutes a dev: XGP => AIXGP GLP => MXGLP DVR => MCDVR DVS => MCDVS TPL => MLTPL Otherwise, it returns "no such device". This means no foreign filesystem can be contacted unless it IS IN THE HARDWIRED TABLE of MLDEV! If is known, then it uses NETWRK"HSTLOOK to find the IP address and proceeds to make a connection. OTHER: Various programs automatically re-init and PDUMP themselves out when they detect that the machine name or system version has changed since the last binary was created. Among these are: SYS;TS NAME SYS;TS PEEK SYS;TS PROBE SYS1;TS METER SYS1;TS PRUFD SYS1;TS SYSMSG SYS2;TS XXFILE SYSBIN;FIDO BIN Other programs that don't automatically repurify themselves, but need to be invoked properly after assembling: SYS;TS MIDAS (sblk orig in SYSBIN;MIDAS *BIN, start at PURIFY) .MAIL.;COMSAT LAUNCH (orig in SYSNET;COMSAT BIN, start at PURIFY) DUMP remembers the name of the machine it was assembled on, and complains if the current system isn't it. Just rebuild. This can be bypassed by setting variable WRONG back to zero after it's started up. DUMP needs SYSENG;FSDEFS > and SYSNET;NETWRK > to assemble. DUMP hacks the file SYSENG;MACRO TAPES. Also hacks .TAPE0; etc?? FTPS is patched so it doesn't try (and fail) to look up its own host name. MAINIT+25/ MOVEI A,[ASCIZ /MIT-NX.US.ORACLE.COM/] ? NOP Patched in SYSBIN;FTPS BIN. Original binary in SYSBIN;FTPS ORGBIN. This patch is unnecessary if the NETWRK hostname lookup software is working (see HOSTS3 below), but the NETHST syscall is buggy cuz it returns the value of IMPUS, not IMPUS3. Before fixing this, have to be sure nothing else will break! Sigh! SYS;NET MAIL - Telnet/supdup server announcement. Revised for NX, original in NET OMAIL. Note: there is a long-standing bug in the TELSER/STY interaction such that the initial %TDORS for a supdup connection happens halfway through a longish greeting message (NET MAIL) and wipes the rest out after the 105th char. Note 23 chars in machine greeting, plus a CRLF or so, to make 130 chars. This may be STY buffer size? SYS;SYSTEM MAIL - universal DDT startup announcement, if any. Flushed, original in SYSTEM OMAIL. SYS;RECORD TIME - Checked & updated to see how long ITS has run. Various programs that check machine name specially: DUMP (as noted above) SYSEN1;PWORD - has table of per-machine specs INQUIR;INQUIR - similar check MACLSP - tries to map predefined atom or something, at UDIRSE. Checks for AI,ML,MC,MD,MX. MLDEV - checks for known machines using 2-char assumption. SYSNET;TELSER - Checks machine name to select greeting header. SYSEN1;SENDER - Checks machine name to see if sending to a known ITS or not. (Doesn't bother to use host table to check sys type!) ITS edits or patches: See also "modlog" which may have more details. FIXED in SYSTEM;ITS 1646 Quantum timer patch: This is an ITS bug that skews the quantum counter timing. UTRNTR+2/ LDB A, [ 125603/ SETZ Q,UPQUAN(U) 044000,,UPQUAN(u) ] FIXED in SYSTEM;IMP 367 NETHST patch: ANETH2+3/ MOVE B,NIFIPA (from MOVEI B,IMPUS) Not a bugfix, just a crude firewall patch: TCPO4+23/ JRST OPNL6 (from CALL PKTGFI) (prevents outbound TCP opens) FIXED in SYSTEM;TIME 951 Y2K patch. Replace: CLCQDT+2/ PUSH P,E-1 PUSH P,E MOVE E-1,FYEAR IDIVI E-1,100. DPB E,[330700,,QDATE] ;YEAR POP P,E POP P,E-1 with: PUSH P,E MOVE E,FYEAR SUBI E,1900. DPB E,[331000,,QDATE] ;YEAR POP P,E NSALV bug: This code in NSALV (part of the INIT startup) is sorta buggy in that when NDRIVE is 1 it is asking for a reverse BLT... SETOM DRIVE ;For now, assume all drives on-line. MOVE A,[DRIVE,,DRIVE+1] BLT A,DRIVE+NDRIVE-1 DSKDMP bug: I forget where this is but am making a note so I remember to someday come back and document it. There is a race condition on booting where DSKDMP initiates a disk transfer that clobbers the code it is about to execute next. It works on a real machine because the CPU is much faster than the disk. Patched, but not really needed now that KLH10 can load ITS directly. DIRECTORY STATUS ================ (NNF = No New Files since final AI/MC snapshot) OK . - Boot load files. See "bootload.files" for more info. Most of old AI stuff flushed as useless. OK .INFO. - NNF, restore all OK .MAIL. - Empty, must fill for COMSAT (and flush old useless files) OK .MSGS. - Empty, old msgs all obsolete & discarded. OK .TAPE0 - Empty is OK; contains some gubbish from KLH dumps. OK .TECO. - NNF, restore all OK .TEMP. - NNF, restore all (empty except for a -READ- -THIS- file) ? C - only partially restored OK CHANNA - NNF, Restore all OK CRASH - Empty. Note COMSAT crashes will leave a BURNUP file here, and each use of CRTSTY adds one line to CRTSTY LOG. OK DEVICE - NNF, Restore all (but not all are needed) However, file CHAOS HOST# isn't present in Alan's snapshot. OK DRAGON - NNF, Restore (but not all are needed; lots of junk!) OK EMACS - NNF, Restore all OK GREN - PERSONAL, NX-only. Flushed from distrib. ? HACK - New, retain for NX, not needed for distrib (but update or remove messages) OK INFO - NNF, Restore all OK INQUIR - NNF, Restore all OK JLV - PERSONAL, NX-only. Flushed from distrib. OK KLH - PERSONAL, NX-only, but may be a few sources that need to be included for distrib: OUT,NUUOS,OUTDOC,MACROS. Move to KSC. (problem; need to update .INSRTs of other progs?) OK KSC - NNF, restore all. ? KSHACK - Mostly new files. Perhaps restore others? Note the DEC ucode source is proprietary; exclude it? OK LISP - NNF, Restore all (needed for INQUIR, sigh) OK MAINT - NNF, Restore all This contains DEC diagnostic binaries but leave it for now. OK MIDAS - NNF except for bugfix: TSRTNS 234 OK MRC - PERSONAL, NX-only. Flushed from distrib. OK SYS - Mostly NNF, restore all. Lots of possibly superfluous large PURQIO files, ugh. But don't know which progs might use which of them... OK SYS1 - NNF, Restore all OK SYS2 - NNF, Restore all OK SYS3 - NNF, Restore all OK SYSBIN - Mostly NNF, restore. Lots of OBINs that could flush. OK SYSDOC - NNF, restore all OK SYSEN1 - NNF, restore all OK SYSEN2 - NNF, restore all OK SYSENG - NNF, restore all except for MACRO TAPES (site-specific) OK SYSHST - Mostly NNF, restore all (new versions of H3TEXT). OK SYSNET - Mostly NNF, restore all (new version of TELSER). OK SYSTEM - Several new files; should clean out old versions. FILE UPDATES AND CHANGES: ======================== This list contains all files added or modified since NX first came up from the original AI/MC snapshot, with descriptive notes. It does NOT contain links; many links have a creation time postdating this but they are almost certainly artifacts of the way link restoral was done. See "nxmodfiles" for a complete list of all files or links that postdate the creation of NX. NX:C; NX:CHANNA; NX:CRASH; 0 CRTSTY LOG 0 +333 ! 9/24/99 18:46:17 (9/24/99) KLH NX:DEVICE; NX:DRAGON; 0 DRAGON HOARD 4 +90 ! 6/11/92 02:56:32 (15/31/127) -??- 0 RAKASH PFTHMG 4 +578 ! 6/11/92 02:56:32 (10/23/99) -??- NX:EMACS; NX:GREN; 0 GREN LOGIN 0 +489 ! 7/23/93 18:50:26 (8/17/93) 0 GREN OSENDS 0 +44 ! 7/23/93 18:55:38 (7/23/93) 0 GREN PLAN 0 +8 ! 7/29/93 12:10:33 (8/17/93) NX:HACK; 0 KLH10 MSG 3 +157 ! 11/8/92 01:46:07 (11/10/92) -??- 0 KLH10 NEWS 0 +660 ! 5/25/93 15:52:15 (5/25/93) KLH 0 NX NEWS 0 +275 ! 5/3/93 14:24:31 (7/22/93) KLH NX:INFO; NX:INQUIR; NX:KLH; 0 COM INFO 0 +229 ! 1/29/93 00:31:09 (5/20/93) 0 CONFIG 1 7 +1 ! 11/7/92 13:41:12 (9/24/99) -??- 0 CONFIG 2 7 +1 ! 9/24/99 04:04:03 (9/24/99) 0 CONFIG 3 6+1016 ! 9/24/99 04:07:38 (9/24/99) 0 DDT NBIN 3 +458 ! 1/5/93 17:15:15 (5/20/93) 0 DDT NEW 3 +449 ! 1/5/93 17:08:04 (5/20/93) 0 DDT TORIG 3 +458 ! 1/5/93 17:08:45 (4/5/93) 0 FIRST EMACS 0 +3 ! 11/8/92 01:29:21 (10/3/99) 0 ITS BIN 56 +346 ! 10/28/93 11:35:05 (10/28/93) 0 KLH MAIL 1 +86 ! 7/23/93 18:45:00 (10/24/99) .MAIL. 0 KLH10 ARCHIV 0 +866 ! 5/25/93 16:00:15 (10/3/99) .MAIL. 0 KLH10 PEOPLE 0 +384 ! 8/5/93 12:29:21 (8/23/93) 0 MAIL FRMHQM 0 +22 ! 11/9/92 20:42:04 (5/20/93) -??- 0 TEMP FILE 0 +13 ! 7/19/92 10:40:14 (11/8/92) 0 WALL 1 5 +152 ! 12/23/92 15:09:31 (12/23/92) 0 WALL 2 0 ! 1/14/93 00:44:45 (1/14/93) 0 WALL 3 8 +15 ! 1/14/93 00:54:11 (1/14/93) 0 WALL 4 8 +15 ! 1/14/93 03:41:59 (1/14/93) 0 WALL 5 0 +133 ! 1/16/93 06:13:50 (1/16/93) 0 WALL 6 7 +992 ! 1/16/93 06:16:51 (1/16/93) 0 WALL 7 7 +992 ! 1/16/93 06:43:49 (1/16/93) 0 XLINKS 1 2 +12 ! 11/8/92 01:20:54 (12/9/92) -??- 0 _FTPU_ OUTPUT 0 ! 1/28/93 17:58:51 (1/28/93) NX:KSC; TS ZOTZ - New, just a ZOTZ BIN patched for 50K (not 500K) loop count 0 TS ZOTZ 3 +317 ! 12/16/92 16:51:33 (9/24/99) KLH NX:KSHACK; 0 NSALV BIN 11 +437 ! 7/20/92 04:35:29 (7/20/92) KLH 0 NXBOOT ALLFS 0 +22 ! 1/16/93 06:16:18 (4/9/93) KLH 0 NXBOOT FILES 0 +42 ! 1/16/93 06:11:07 (4/9/93) KLH NX:MAINT; NX:MIDAS; 0 TS NMIDAS 22 +357 ! 6/23/92 15:10:28 (4/9/93) -??- 0 TSRTNS 234 27 +612 ! 6/23/92 17:23:37 (4/9/93) -??- NX:MRC; NX:SYS; NET MAIL - Telnet/supdup server announcement. Revised for NX, original in NET OMAIL. :MSGS TIMES - Used by DDT's :MSGS command to remember the last time each user invoked it; data file of uname/datim pairs. TS MIDAS - New, version 458. Old version (433) renamed to TS OMIDAS. 0 :MSGS TIMES 4 ! 10/24/99 01:53:05 (10/24/99) KLH 0 NET MAIL 0 +16 ! 5/3/93 14:28:10 (10/24/99) KLH 0 RECORD TIME 0 ! 6/19/92 17:19:49 (9/24/99) KLH 0 TS MIDAS 22 +357 ! 4/9/93 21:36:42 (9/24/99) KLH 0 TS NAME 16 +108 ! 4/15/93 19:07:13 (10/3/99) KLH 0 TS PEEK 14 +726 ! 4/7/93 19:03:07 (10/24/99) KLH 0 TS PROBE 7 +8 ! 7/23/93 18:54:38 (7/23/93) GREN NX:SYS1; 0 TS METER 3 +398 ! 5/19/93 19:12:17 (5/20/93) MRC 0 TS PRUFD 0 +506 ! 4/6/93 17:23:25 (4/6/93) KLH 0 TS SYSMSG 2 +568 ! 6/7/93 13:04:43 (6/21/93) KLH NX:SYS2; 0 TS XXFILE 5 +982 ! 4/5/93 19:59:23 (4/5/93) KLH NX:SYS3; NX:SYSBIN; FTPS BIN Patched with new IP address; orig is FTPS ORGBIN. MIDAS 458BIN New version, latest. 0 FIDO BIN 3 +430 ! 5/11/93 13:39:52 (5/11/93) KLH 0 FTPS BIN 10 +367 ! 11/7/92 13:40:41 (5/19/93) KLH 0 HOSTS3 2002 68 +757 ! 5/19/93 12:51:50 (10/24/99) KLH 0 MIDAS 458BIN 17 +108 ! 4/9/93 21:26:11 (4/9/93) KLH 0 TELSER BIN 3 +87 ! 5/3/93 13:53:46 (10/24/99) KLH New version (TELSER 173) with NX greeting. Old file renamed to TELSER AIBIN. NX:SYSDOC; NX:SYSEN1; NX:SYSEN2; NX:SYSENG; Must flush this site-specific weirdo; it will conflict with version from dump! DUMP maintains this as a directory of all macro tapes written on the system. 0 MACRO TAPES 40 ! 3/1/93 21:43:22 0 MACRO TAPES 39 +64 ! 9/24/99 04:32:12 (9/24/99) KLH NX:SYSHST; New versions of H3TEXT >. 0 H3TEXT 2000 82 +412 ! 5/17/93 18:36:38 (5/17/93) -??- 0 H3TEXT 2002 74 +35 ! 5/17/93 19:33:08 (5/19/93) KLH NX:SYSNET; 0 TELSER 173 7 +886 ! 5/3/93 13:53:31 (5/3/93) KLH New version with NX greeting. NX:SYSTEM; Several new files; should clean out old versions. 0 CHAOS 289 16 +328 ! 7/19/92 09:31:21 (7/19/92) KLH 0 CONFIG 197 6+1008 ! 7/19/92 11:00:02 (9/24/99) KLH 0 CONFIG 198 6+1016 ! 9/24/99 04:07:38 (9/24/99) KLH 0 DDT BIN 3 +671 ! 7/19/92 13:53:44 (7/19/92) KLH 0 DSKDMP 216 9 +809 ! 7/19/92 13:56:53 (4/21/93) KLH 0 DSKDMP BIN 2 +296 ! 7/20/92 04:35:00 (7/20/92) KLH 0 DSKDMP BT 0 +885 ! 7/20/92 04:35:16 (7/20/92) KLH 0 IMP 366 6 +216 ! 7/20/92 04:20:41 (9/24/99) KLH 0 INET 139 9 +838 ! 7/20/92 04:58:30 (9/24/99) KLH 0 ITS 1645 118 +583 ! 7/19/92 06:12:26 (9/24/99) KLH 0 ITS 645BIN 56 +345 ! 7/20/92 04:30:13 (11/23/92) KLH 0 ITS 645NX 56 +346 ! 9/24/99 04:09:48 (9/24/99) KLH 0 ITS BIN 56 +346 ! 9/24/99 19:00:59 (9/24/99) KLH 0 ITS OBIN 56 +345 ! 1/28/93 00:16:44 (1/28/93) KLH 0 NET 33 2 +725 ! 7/19/92 06:06:21 (9/24/99) KLH 0 TM03S DEFS5 1 +65 ! 7/20/92 04:23:56 (9/24/99) KLH 0 TTYTYP 315 3 +69 ! 7/19/92 11:45:53 (9/24/99) KLH NX:.; 0 @ AI-ITS 58 +29 ! 12/17/92 17:04:48 (4/9/93) -??- 0 @ ITS 62 +464 ! 12/17/92 17:05:37 (1/14/93) -??- 0 @ NSALV 12 +675 ! 12/17/92 17:05:03 (4/19/93) -??- 0 @ SALV38 12 +540 ! 6/4/92 12:32:09 (1/14/93) -??- 0 DDT 068BIN 3 +671 ! 7/19/92 13:53:44 (4/9/93) -??- 0 DSKDMP 216BIN 2 +296 ! 7/20/92 04:35:00 (4/9/93) -??- 0 DSKDMP 216BT 0 +885 ! 7/20/92 04:35:16 (4/9/93) -??- 0 ITS 645BIN 52 +240 ! 7/20/92 04:34:15 (4/9/93) -??- 0 NSALV 260BIN 11 +437 ! 7/20/92 04:35:29 (4/19/93) -??- NX:.INFO.; NX:.MAIL.; Empty until COMSAT launched for first time. Most of the old files were useless, but some saved for posterity. 0 ID 14 0 ! 5/19/93 18:45:58 (7/23/93) KLH 0 LIST EQV 2 +191 ! 5/24/93 15:58:39 (7/23/93) 0 LIST QUEUE 0 +203 ! 5/24/93 15:58:39 (5/24/93) 0 LIST MASTER 0 +203 ! 5/24/93 15:58:39 (8/20/93) 0 LIST REMIND 0 +201 ! 5/19/93 18:45:58 (5/24/93) KLH 0 LISTS MSGS 0 ! 5/19/93 18:45:58 (5/24/93) KLH 0 COMSAT LAUNCH 32 +463 ! 5/19/93 18:45:44 (10/23/99) KLH 0 NAMED ERR000 0 +8 ! 5/19/93 18:46:00 (5/19/93) KLH 0 NAMED ERR001 0 +8 ! 5/24/93 15:58:39 (5/24/93) 0 NAMES 2000 1 +32 ! 5/19/93 17:03:19 (5/24/93) KLH 0 NAMES 2001 1 +59 ! 5/24/93 15:56:18 (8/20/93) KLH 0 STATS 1 1 +738 ! 10/23/99 00:29:02 (10/23/99) -??- 0 XLIST EQV 2 +137 ! 5/19/93 17:04:49 (5/19/93) KLH 0 XLIST MSGS 0 +201 ! 5/19/93 17:07:49 (5/19/93) 0 XLIST QUEUE 0 +201 ! 5/19/93 17:23:04 (5/19/93) KLH 0 XLIST MASTER 0 +203 ! 5/19/93 17:23:04 (5/19/93) KLH 0 XLIST REMIND 0 +201 ! 5/19/93 17:04:47 (5/19/93) KLH 0 XNAMED ERR064 0 +9 ! 5/19/93 16:13:18 (5/19/93) KLH 0 XNAMED ERR065 0 +9 ! 5/19/93 17:04:49 (5/19/93) KLH L XNAMES 1062 .MAIL.;NAMES 1062 5/17/93 14:55:27 (5/17/93) KLH 0 XNAMES 1064 0 +391 ! 5/17/93 16:53:36 (5/19/93) KLH 0 XNAMES 1065 1 +32 ! 5/19/93 17:03:19 (5/19/93) KLH 0 XSTATS 1 3 +166 ! 5/19/93 18:31:57 (5/19/93) KLH NX:.TAPE0; L LAST USE KLH;LIST 990924 9/24/99 04:32:19 (9/24/99) KLH 0 TAPE 0 0 +46 ! 9/24/99 04:32:13 (9/24/99) KLH 0 TAPE 1 0 +10 ! 4/19/93 18:47:27 (4/19/93) -??- 0 TAPE 2 3 +687 ! 4/19/93 18:35:28 (4/19/93) -??- 0 TAPE 3 0 +484 ! 4/21/93 21:07:36 (4/21/93) -??- NX:.TECO.; NX:.TEMP.; Note fucking DUMP gubbish to do FS save/restore properly: To save stuff, use: DUMP FULL LINKS To restore stuff, use: RELOAD LINKS CRDIR SORRY ??? (Don't use LOAD, use RELOAD!!!) Fucking DUMP also appears to lose when loading up a new dir when the first file is a link, even if the CRDIR option is set!!! Need to create the dir by hand, then continue. Idea for new DUMP format to preserve stuff better: Extend file header to 9 words. Move Dump-check and don't-reap bits to LH of link,,pack# Add sixbit author name at end. Must modify: NSALV - to handle bits and potentially convert author to index. Potentially store MFD if found, use to generate new UFDs as needed. DUMP - to output and read in extended format. ITS - add new syscall to read/write bytesize and file length? RFSIZE, SFSIZE arg1 - bytesize, arg2 - # bytes For SFSIZE, # bytes must be within last word or at start of next, else error. if (BS <= 0 || 36 < BS) error; BPW = 36/BS; if (((# + BPW-1) / BPW) != FILLEN) error; Date: Wed, 21 Apr 93 19:10:33 -0400 From: Alan Bawden To: klh@us Subject: ITS Turist Tips Date: Wed, 21 Apr 1993 9:34:08 BST From: Ken Harrenstien Sure, but why not run it? The statistics are sometimes interesting. Also PFTHMG runs hourly, daily, monthly and yearly batch jobs, which is a useful service in general... I guess. Now you have to explain what those various batch jobs are doing there... Let's see... (Wow, there seem to be more of these than I remember!) On AI, hourly we have: DRAGON;HOURLY CNAVRL ==> .MAIL.;COMSAT LAUNCH Just makes sure COMSAT doesn't stay dead for long. DRAGON;HOURLY GCMAIL This is something SRA wrote to garbage collect files from the .MAIL. directory. I don't remember exactly what it does -- perhaps it deletes OSTATS files that are older than a few days or something. I do remember that there was some disagreement about how aggressive this should be. Penny didn't want it as aggressive as SRA did. Penny won. DRAGON;HOURLY H3GET This is something I wrote that keeps the host tables on all the ITS machines in sync. It checks the host tables on other ITS machines, and if it finds one that is newer than the local copy, it sucks it over. DRAGON;HOURLY MODEMS ==> CHANNA;TS MODEMS This does the same thing as CHANNA;RAKASH MODEMS, this way the damn things only stay broken for at most an hour. DRAGON;HOURLY TMPKIL ==> SYS2;TS TMPKIL I think this has something to do with cleaning up the files on .TEMP. -- I don't know its algorithm. DRAGON;HOURLY UPTIME ==> BAWDEN;UPTIME BIN This is just a little personal hack I wrote to keep track of system uptime. It just stores the current system uptime in the creation date of some file on my directory. Then every night soon after midnight we run: DRAGON;DAILY BDAY This mails out the daily birthday greetings mail. DRAGON;DAILY EXPIRE ==> SYS2;TS GMSGS This cleans out the expired system messages on .MSGS.. DRAGON;DAILY NO.TS This searches for files named GUEST*;TS * and USER*;TS * and deletes them. (People sharing a directory shouldn't create TS files for each other to trip over by accident.) On the first of the month, and the first of the year we run: DRAGON;MNTHLY MNTHLY ==> DRAGON;YEARLY YEARLY DRAGON;YEARLY YEARLY This just sends mail to Alan to let him know that Puff has actually run the monthly and yearly batch jobs, even though normally these are the only entries. On MC, we also have HOURLY DIGEST, which is the mailing list digestifier (and maybe some others I don't remember). And what's PANDA? RWK;TS PANDA is the administrator's interface to PWORD. Its lets you turn people's accounts on and off and stuff like that. Since you aren't going to run PWORD, you don't need it. Host table madness: Several programs will die on startup if they cannot find their own host address or look up their name using this address. Amongst them are FTPS. SYSHST;H3MAKE > - source for a small program that DRAGON is supposed to run periodically. It invokes SYSHST;HOSTS3 BIN (compiled from SYSHST;HOSTS3 >) with appropriate non-trivial args that cause it to snarf up the file SYSHST;H3TEXT > and compile it into the binary-format file SYSBIN;HOSTS3 > which is used by everything else on ITS to look up host names via this package: SYSNET;NETWRK > Note that COMSAT uses a new roundabout mechanism. Instead of the NETWRK library it now uses SYSNET;RESOLV > which replaces all NETWRK calls and invokes the DQ device, installed as DEVICE;JOBDEV DQ compiled from SYSNET;DQXDEV > which itself uses NETWRK to look up HOSTS3 information. This DQ device was intended to be a temporary but more reliable replacement for SYSNET;DQDEV > which is a CSTACY product attempting to do real DNS lookups via UDP. MIT-NX changes: A new H3TEXT > was pulled over from lcs.mit.edu:/hosts/.misc/hosts3.txt This file is generated automatically by a batch job. It was edited to remove X-terminals and chaosnet hosts, and some NX-local names added. COMSAT configuration: Must eventually flush its HN$ symbols. At BUGHST/ must put the IP address of the host that deals with (BUG RANDOM-PROGRAM) messages. At DOMGAT/ must put the IP address of a mail relay host. This host will receive all mail for addresses that we can't resolve ourselves. At TCPGAT/ can put IP address of a mail relay to handle *all* external TCP/IP traffic. This is needed when there is a firewall host like gatekeeper.us.oracle.com, sigh. NX .MAIL. FREE BLOCKS #0=18478 0 ID 703539 0 ! 4/23/85 10:27:26 FLUSHED 0 ID 723380 0 ! 4/23/85 10:27:26 FLUSHED 0 LIST EQV 15 ! 4/19/90 17:41:08 FLUSHED 0 LIST QUEUE 2 ! 4/20/90 11:07:06 FLUSHED 0 LIST MASTER 5 ! 4/20/90 11:05:48 FLUSHED 0 LIST REMIND 1 ! 11/19/85 18:46:42 FLUSHED 0 LISTS MSGS 361 ! 4/20/90 11:05:47 FLUSHED 0 LOCK UNIQUE 1 ! 4/23/85 10:27:24 0 *MSG EXP 2 ! 1/30/90 20:26:07 0 -READ- -THIS- 1 ! 12/19/86 13:47:38 0 .AIOLD NAMES 17 ! 7/27/85 16:35:38 0 .DM NAMES 3 ! 7/22/83 13:31:25 0 .MCNEW NAMES 19 ! 5/17/86 00:45:39 0 .MCOLD NAMES 24 ! 5/5/86 13:41:19 0 .MLOLD NAMES 9 ! 2/21/84 00:53:15 0 BBOARD INFO 1 ! 4/29/84 18:01:28 0 COMSAT BIN 31 ! 12/23/89 16:01:35 0 COMSAT LAUNCH Patched with: BUGHST/ self DOMGAT,TCPGAT/gateway 0 COMSAT LUNCH 33 ! 9/19/90 16:17:58 AI Original 0 COMSAT OBIN 28 ! 11/10/89 17:55:53 0 COMSAT OLAUNC 33 ! 11/10/89 17:56:40 FLUSHED (pure version of OBIN) 0 FAILED STUFF 24 ! 3/29/90 22:36:45 FLUSHED 0 MAIL 1 1 ! 9/20/90 18:15:55 FLUSHED (random ALAN stuff) 0 MAIL 2 1 ! 9/20/90 18:15:55 FLUSHED 0 MAILIN 1 1 ! 4/20/90 11:16:56 FLUSHED 0 NAMED ERR094 1 ! 4/5/90 22:19:22 FLUSHED 0 NAMED ERR096 1 ! 4/14/90 02:10:28 FLUSHED 0 NAMED ERR097 1 ! 4/19/90 17:41:08 FLUSHED 0 NAMES 1096 7 ! 4/14/90 02:06:27 FLUSHED 0 NAMES 1097 7 ! 4/19/90 17:36:00 Renamed to .AIKS NAMES 0 NAMES INFO 3 ! 7/24/89 23:41:25 0 OSTATS 1634 51 ! 4/18/90 16:31:55 FLUSHED 0 OSTATS 1635 51 ! 4/19/90 23:07:52 FLUSHED 0 XID 22 0 ! 10/30/89 19:12:43 FLUSHED 0 XLIST EQV 1 ! 10/30/89 22:30:30 FLUSHED 0 XLIST MSGS 0 ! 10/31/89 01:40:09 FLUSHED 0 XLIST QUEUE 1 ! 12/23/89 16:08:06 FLUSHED 0 XLIST MASTER 1 ! 12/23/89 16:08:06 FLUSHED 0 XLIST REMIND 1 ! 10/30/89 19:18:24 FLUSHED 0 XLOCK UNIQUE 1 ! 10/30/89 19:12:40 FLUSHED 0 XNAMED ERR062 1 ! 10/30/89 19:20:18 FLUSHED 0 XNAMED ERR063 1 ! 10/30/89 22:30:29 L XNAMES 1062 .MAIL. NAMES 1062 0 XNAMES 1063 1 ! 10/30/89 22:23:34 0 XSTATS 1 3 ! 12/23/89 16:18:39 NX SYSBIN FREE BLOCKS #0=18565 0 BIG 0DAT 15 ! 4/20/90 11:26:41 L -READ- -THIS- SYSENG -READ- -THIS- 0 AP BIN 1 ! 3/13/84 05:25:26 0 BOLIO 174 63 ! 12/13/82 20:16:15 0 BOLIO 177 67 ! 12/7/83 03:56:14 0 BOLIOP 191.1 1 ! 1/25/83 03:34:24 0 BOLIOP 191.2 1 ! 4/7/83 00:20:25 0 BOLIOP 191.3 1 ! 6/7/83 20:05:15 0 BOLIOP 191PAT 1 ! 6/7/83 20:06:01 0 BOLIOP 195PAT 1 ! 12/7/83 03:55:56 0 BRANDX 45 58 ! 12/20/82 18:05:01 0 BX45 (PDIR) 1 ! 12/20/82 18:04:59 0 BXL23 (PDIR) 1 ! 8/31/81 01:52:02 0 BXLSB 30 92 ! 12/20/82 18:20:08 0 CHTN BIN 4 ! 10/22/86 22:02:55 0 CNVRT BIN 2 ! 9/29/76 02:35:32 0 CRTSTY BIN 30 ! 8/2/89 20:58:34 0 CRTSTY OBIN 27 ! 7/22/85 13:18:43 0 CTN BIN 28 ! 12/31/84 11:07:05 0 CTN OBIN 28 ! 12/31/83 08:42:22 0 CTN SCRTST 25 ! 3/17/82 04:28:57 0 DDT 545BIN 27 ! 11/26/88 16:04:38 0 DUMP 420BIN 14 ! 6/2/88 17:59:35 0 DUMP 422BIN 14 ! 9/20/88 14:46:06 0 DUMP 442BIN 14 ! 8/4/89 15:17:23 L DUMP BIN SYSBIN DUMP > 0 FIDO BIN 4 ! 5/11/93 13:39:52 0 FILE BIN 8 ! 12/19/86 17:12:14 0 FILE NBIN 8 ! 2/20/89 18:08:25 0 FTPS BIN 11 ! 11/7/92 13:40:41 0 FTPS OBIN 12 ! 8/2/89 21:16:07 0 FTPS ORGBIN 11 ! 11/27/89 00:33:06 0 FTPU BIN 11 ! 8/2/89 21:12:42 0 FTPU OBIN 11 ! 3/5/89 15:00:03 0 GMSGS OBIN 3 ! 2/17/89 03:02:49 0 HOST1 OLD548 3 ! 3/15/83 19:00:48 0 HOSTS2 29 23 ! 5/2/85 19:22:53 0 HOSTS2 31 24 ! 6/16/85 01:54:29 0 HOSTS3 1560 72 ! 4/16/90 05:44:54 FLUSHED 0 HOSTS3 1561 72 ! 4/18/90 05:16:33 FLUSHED 0 HOSTS3 1562 72 ! 4/20/90 05:31:08 Renamed to HOSTS3 AI1562 0 HOSTS3 OLD 82 ! 2/24/89 04:48:23 0 HOSTS3 OLDOLD 71 ! 4/21/87 19:51:20 0 HST3BD 2 53 ! 8/23/85 01:48:56