SSeettttiinngg UUpp UUnniixx -- SSeevveenntthh EEddiittiioonn _C_h_a_r_l_e_s _B_. _H_a_l_e_y _D_e_n_n_i_s _M_. _R_i_t_c_h_i_e This document has been updated to include additional processor and peripheral device support , and other enhance- ments incorporated in UNIX(R) version seven by the Unix Sys- tems Engineering group at DIGITAL. The addenda tape (dated 12/2/80) supplied with the standard Unix version seven dis- tribution has been included on this tape in the directory /sys/addenda. Most of the addenda has already been installed on this distribution, for more detail about the addenda see /sys/addenda/README. Unix/v7m Release 2.1 Fred Canter Bill Shannon Jerry Brenner Armando Stettner Digital Equipment Corporation assumes no responsibili- ties for this software and makes no warranties or guaranties as to its suitability or completeness. -2- If you are set up to do it, it might be a good idea immediately to make a copy of the tape to guard against dis- aster. The tape is 9-track 800 BPI or 1600 BPI and contains some 512-byte records followed by many 10240-byte records. There are interspersed tapemarks. The system as distributed contains binary images of the system and all the user level programs, along with source and manual sections for them--about 2100 files altogether. The binary images, along with other things needed to flesh out the file system enough so UNIX will run, are to be put on one file system called the `root file system'. The file system size required is 9600 blocks. The second file system contains the system sources, needed to rebuild Unix and requires about 5000 512-byte blocks. The third file system has the remainder of the sources and all of the documenta- tion, it requires 20,000 blocks of storage. MMaakkiinngg aa DDiisskk FFrroomm TTaappee Perform the following bootstrap procedure to obtain a disk with a root file system on it. 1. Mount the magtape on drive 0 at load point. 2. Mount a formatted disk pack on drive 0. In some cases a disk pack must also be mounted on drive 1. 3. Use the DEC hardware boot ROM or other means to load block 0 or 1 at 800 or 1600 BPI into location 0 and transfer to 0. 4. The tape should move and the CPU loop. 5. The console should type _B_o_o_t _: Copy the magtape to disk by the following procedure. The machine's printouts are shown in " ", explanatory comments are within ( ). Terminate each line you type by carriage return or line-feed. There are three classes of tape drives: the name `tm' is used for the TU10 or TE10, `ts' is used for the TS11, and `ht' is used for the TU16 or TE16. There are also four classes of disks: `rp' is used for the RP03, `hp' is used for the RM02/3 and the RP04/5/6, `hk' is used for the RK06/7 , and `rl' is used for the RL02. -3- If you should make a mistake while typing, the charac- ter '#' erases the last character typed up to the beginning of the line, and the character '@' erases the entire line typed. Some consoles cannot print lower case letters, adjust the instructions accordingly. (bring in the program mkfs) ":" tm(0,3) (`tm(0,3)' for TU10/TE10 800 BPI) (`ht(0,3)' for TU16/TE16 800 BPI) (`ht(4,3)' for TU16/TE16 1600 BPI) (`ts(0,3)' for TS11 1600 BPI) (`ts(6,3)' for TS11 1600 BPI `tstm') "file system size:" 9600 (9600 for RP03) (9600 for RM02/3, RP04/5/6) (9600 for RK06/7) (7240 for RL01) (16480 for RL02) (see Table 1, in Disk Layout below, for values of `m' and `n') "interleave factor:" m "blocks per cylinder:" n "file system:" rp(0,0) (`rp(0,0)' for RP03) (`hp(0,0)' for RM02/3, RP04/5/6) (`hm(0,0)' for hp on second RH) (`hk(0,0)' for RK06/7) (`rl(0,0)' for RL01/2) _"_i_s_i_z_e _= _X_X_" _"_m_/_n _= _X_X_" (after a while) _"_e_x_i_t _c_a_l_l_e_d_" _"_B_o_o_t_" _"_:_" The above step makes an empty file system. -4- 6. The next thing to do is to restore the data onto the new empty file system. To do this you respond to the `:' printed in the last step with (bring in the program restor) ":" tm(0,4) (`tm(0,4)' for TU10/TE10 800 BPI) (`ht(0,4)' for TU16/TE16 800 BPI) (`ht(4,4)' for TU16/TE16 1600 BPI) (`ts(0,4)' for TS11 1600 BPI) (`ts(6,4)' for TS11 1600 BPI `tstm') "tape?" tm(0,5) (`tm(0,5)' for TU10/TE10 800 BPI) (`ht(0,5)' for TU16/TE16 800 BPI) (`ht(4,5)' for TU16/TE16 1600 BPI) (`ts(0,5)' for TS11 1600 BPI) (`ts(6,5)' for TS11 1600 BPI `tstm') "disk?" rp(0,0) (`rp(0,0)' for RP03) (`hp(0,0)' for RM02/3, RP04/5/6) (`hm(0,0)' for hp on second RH) (`hk(0,0)' for RK06/7) (`rl(0,0)' for RL01/2) "Last chance before scribbling on disk." (you type return) (the tape moves, perhaps 5-10 minutes pass) "end of tape" _"_B_o_o_t_" _"_:_" You now have a UNIX root file system. -5- BBoooottiinngg UUNNIIXX You probably have the bootstrap running, left over from the last step above; if not, repeat the boot process (step 3) again. The root file system contains 18 preconfigured unix operating systems. These unix systems are named `xyu- nix', where `x' is the disk name and `y' is the tape name. Boot the unix monitor which most closely matches your con- figuration by responding to the ":" with one of the follow- ing: hp(0,0)hphtunix (for RP04/5/6 or RM02/3,) hp(0,0)hptmunix (on first RH controller) hp(0,0)hptsunix hm(0,0)hmhtunix (for RP04/5/6 or RM02/3,) hm(0,0)hmtmunix (on second RH controller) hm(0,0)hmtsunix rp(0,0)rphtunix (for RP03) rp(0,0)rptmunix rp(0,0)rptsunix hk(0,0)hkhtunix (for RK06/7) hk(0,0)hktmunix hk(0,0)hktsunix rl(0,0)rl01htunix (for RL01) rl(0,0)rl01tmunix rl(0,0)rl01tsunix rl(0,0)rl02htunix (for RL02) rl(0,0)rl02tmunix rl(0,0)rl02tsunix The machine should type the following: _u_n_i_x_/_v_7_m _2_._1 _m_e_m _= _x_x_x _# The _m_e_m message gives the memory available to user programs in bytes. After booting unix, the current date and time (date (1)) must be set as follows: # date yymmddhhmm i.e. # date 8108231300 -6- UNIX is now running, and the `UNIX Programmer's manual' applies; references below of the form X(Y) mean the subsec- tion named X in section Y of the manual. If your processor is one of the non-separate I & D space CPUs mentioned above, you must use the alternate version of several system com- mands and the C object code optimizer. If required the alternate commands should be installed as follows: cd /bin cat makefile make cmd40 (non-separate I & D CPU) or make cmd70 (separate I & D CPU) cd / sync The system is distributed with the `cmd70' commands installed. The following commands will be replaced: adb dcheck dump dumpdir icheck ncheck restor tar The above list represents the commands required to set up and maintain unix version seven on the smaller processors. Most of the remaining commands will function without being recompiled, however there will be certain commands that will need to be recompiled and others that are too large to oper- ate on non-separate I & D space CPUs. If your system disk is the RL01 or RL02, you must select alternate `/etc/rc' and `/etc/checklist' files as follows: cp /etc/rc_rl01 /etc/rc cp /etc/cl_rl01 /etc/checklist or cp /etc/rc_rl02 /etc/rc cp /etc/cl_rl02 /etc/checklist The `#' is the prompt from the Shell, and indicates you are the super-user. The user name of the super-user is `root' if you should find yourself in multi-user mode and need to log in; the login is `root', there is no password. -7- To simplify your life later, copy the appropriate ver- sion of the system as specified above to plain `unix.' For example, use cp (1) as follows if you have an RP03 disk and a TU16/TE16 tape: cp rphtunix unix In the future, when you reboot, you can type just hp(0,0)unix to the `:' prompt. (Choose appropriately among `hp', `rp', `hk', `rl' , `ht', `tm', `ts' according to your configura- tion). After selecting and copying the correct version of unix, all of the remaining preconfigured unix systems should be removed in order to free up space in the root file sys- tem. The appropriate preconfigured unix system should be retained so that in case of disaster there will always be a bootable unix system in the root. You now need to make some special file entries in the dev directory. These specify what sort of disk you are run- ning on, what sort of tape drive you have, and where the file systems are. For simplicity, this recipe creates fixed device names. These names will be used below, and some of them are built into various programs, so they are most con- venient. However, the names do not always represent the actual major and minor device in the manner suggested in section 4 of the Programmer's Manual. For example, `rp3' will be used for the name of the file system on which the user file system is put, even though it might be on an RP06 and is not necessarily logical device 3. Also, this sequence will put the user file system on the same disk drive as the root, which is not the best place if you have more than one drive. Thus the prescription below should be taken only as one example of where to put things. See also the section on `Disk layout' below. -8- In any event, change to the dev directory (cd(1)) and, if you like, examine and perhaps change the makefile there (make (1)). cd /dev cat makefile Then, use one of make rp03 make rp04 make rp05 make rp06 make rm02 make rm03 make rp04_2 make rp05_2 make rp06_2 make rm02_2 make rm03_2 make rk06 make rk07 make rl01 make rl02 depending on which disk you have. Then, use one of make tm make ht make ts make tstm depending on which tape you have. The `_2' lines are for the hp disks on the second RH11 or RH70 controller. The `tstm' line is for the special case of having a TS11 and a TM11 on the same system. The file `rp0' refers to the root file system; `swap' to the swap-space file system; `rp2' to the system source file system; 'rp3' to the user file system. The devices `rrp0', 'rrp2', and `rrp3' are the `raw' versions of the disks. Also, `mt0' is tape drive 0, at 800 BPI; `rmt0' is the raw tape, on which large records can be read and writ- ten; `nrmt0' is raw tape with the quirk that it does not rewind on close, which is a subterfuge that permits multi- file tapes to be handled. `mt1', `rmt1', and `nrmt1' also refer to tape drive 0, but at 1600 BPI. -9- Before anything further is done the bootstrap block on the disk (block 0) should be filled in. This is done using one of the following commands: (RP03) dd if=/mdec/rpuboot of=/dev/rp0 count=1 (RM02/3, RP04/5/6) dd if=/mdec/hpuboot of=/dev/rp0 count=1 (RK06/7) dd if=/mdec/hkuboot of=/dev/rp0 count=1 (RL01/2) dd if=/mdec/rluboot of=/dev/rp0 count=1 Now the DEC disk bootstraps are usable. See Boot Proce- dures(8) for further information. The system sources should now be extracted from the distribution tape by following the procedure shown below, again the comments are enclosed in ( ). The number follow- ing the "mkfs" command is the size of the file system, which differs for each type of disk. The numbers `m' and `n', following the size , are the file system interleave factors, see `Disk Layout' table 1 to obtain the optimum values for your system. (make an empty file system) /etc/mkfs /dev/rrp2 ##### m n (skip this if using RL02) (07400 on RP03) (08000 on RM02/3) (08778 on RP04/5/6) (08514 on RK06/7) (10240 on RL01) /etc/mount /dev/rp2 /sys (skip this if using RL02) cd /sys (skip 6 files on the tape) dd if=/dev/nrmt0 of=/dev/null bs=20b files=6 ( 800 BPI) dd if=/dev/nrmt1 of=/dev/null bs=20b files=6 (1600 BPI) (extract the source files) tar40 xbf 20 /dev/rmt0 ( 800 BPI) tar40 xbf 20 /dev/rmt1 (1600 BPI) cd / sync /etc/umount /dev/rp2 (skip this step if using RL02) -10- The next thing to do is to extract the rest of the data from the tape. This user file system is too large for a single RL01 disk, it must be split between two RL01 disk packs. There is no set procedure for accomplishing the split, one suggestion would be to load the source files on one pack, the documentation on the second pack and then fill out each pack with the remainder of the files. Comments are enclosed in ( ); don't type these. The number in the first command is the size of the file system; it differs between RP03, RM02/3, RP04/5, RP06 , RK06/7, and RL01/2. Again see `Disk Layout' table 1 for the values of `m' and `n'. (make an empty file system) /etc/mkfs /dev/rrp3 ###### m n (055000 on RP03) (105120 on RM02/3) (144210 on RP04/5) (313082 on RP06) (010240 on RL01) (020480 on RL02) (027060 on RK06) (026598 on RK07) (The above command takes about 2-3 minutes on an RP03) /etc/mount /dev/rp3 /usr cd /usr (skip 7 files on the tape) dd if=/dev/nrmt0 of=/dev/null bs=20b files=7 ( 800 BPI) dd if=/dev/nrmt1 of=/dev/null bs=20b files=7 (1600 BPI) (extract the files) tar40 xbf 20 /dev/rmt0 ( 800 BPI) tar40 xbf 20 /dev/rmt1 (1600 BPI) (This takes a while, time for a break !) cd / sync /etc/umount /dev/rp3 All of the data on the tape has been extracted. You may at this point mount the source file systems (mount(1)). To do this type the following: /etc/mount /dev/rp2 /sys (skip this step if using RL02) /etc/mount /dev/rp3 /usr The source and manual pages are now available in subdirecto- ries of /usr and /sys. The above mount commands are only needed if you intend to play around with source on a single user system, which is not necessary. The file systems are mounted automatically when multi-user mode is entered, by a command in the file /etc/rc. (See `Disk Layout' below). The system should now be placed into multi-user mode by typing the control and d keys simultaneously (control d), some text will be printed followed by the "login:" prompt. -11- Before UNIX is turned up completely, a few configura- tion dependent exercises must be performed. At this point, it would be wise to read all of the manuals (especially `Regenerating System Software') and to augment this reading with hand to hand combat. DO NOT !, proceed past this point until you have read the section on regenerating UNIX in `Regenerating System Software'. The `Setting Up Unix' and the `Regenerating System Software' documents are located in /usr/doc and can be printed by using one of the following commands: nroff -mf setup nroff -mf regen RReeccoonnffiigguurraattiioonn The general information given in this section is intended to document the areas of the unix software which require modification in order to tailor unix to the specific system configuration. The suggested procedure is to read this section and then follow the step by step instructions in the `UNIX' section of `Regenerating system Software'. The first step is to select the type of unix system most appropriate to your CPU, there are three; unix_id for the separate I & D space CPUs (PDP11/44, PDP11/45, PDP11/55, & PDP11/70); unix_i for the non-separate I & D space CPUs (PDP11/23, PDP11/24, PDP11/34, PDP11/40, & PDP11/60); unix_ov the overlay text kernel for the non-separate I & D space CPUs. The size limitations of the unix_i kernel makes its use as an actual unix system inappropriate, its only purpose in life is to initially load unix from the distribu- tion tape. The overlay text unix kernel should be used as the multi-user unix system on the non-separate I & D space machines. The UNIX system running is configured to run on a non- separate I & D space CPU with the given disk and tape, a console, and no other device. This is certainly not the correct configuration. You will have to correct the config- uration table to reflect the true state of your machine. -12- It is wise at this point to know how to recompile the system. Respond to the "login:" prompt with "sys" followed by a return, there is no password. This will log you into the system source account with a current directory of /sys. Print (cat(1)) the file /sys/conf/makefile. This file is input to the program `make(1)' which if invoked with `make all' will recompile all of the system source and install it in the correct libraries. The libraries supplied with this distribution tape are up to date and need not be recompiled unless changes are made to the system source code. See `Regenerating System Software' for instructions on recompil- ing individual source modules and installing them in the correct libraries. The file /sys/h/param.h contains the parameters which determine the size of the various data spaces within unix. There are two versions of this file, /sys/h/param_ov.h for the PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60, and /sys/h/param_id.h for the PDP11/44, PDP11/45, and PDP11/70. Prior to recompiling any of the system source code the correct parameter file must be copied to /sys/h/param.h and to /usr/include/sys/param.h. This is done automatically by the makefiles and shell procedues which are provided for making unix. The program mkconf(1) prepares files that describe a given configuration (See mkconf(1)). In the /sys/conf directory, the 18 files xyconf were input to mkconf to pro- duce the 18 versions of the system xyunix, "x" is the disk type (rp, hp, hm, hk, rl01, or rl02) and "y" is the tape (ht, tm, or ts). Pick the appropriate one, copy it to unix- conf, and edit unixconf to add lines describing your own configuration. (Remember the console typewriter is automat- ically included; don't count it in the kl specification.) Then run mkconf; it will generate the files l.s (trap vec- tors) c.c (configuration table), and mch0.s. Take a careful look at l.s to make sure that all the devices that you have are assembled in the correct interrupt vectors. If your configuration is non-standard, you will have to modify l.s to fit your configuration. -13- There are certain magic numbers and configuration parameters imbedded in various device drivers that you may want to change. The device addresses of each device are defined in each driver. In case you have any non-standard device addresses, just change the address and recompile. (The device drivers are in the directory /sys/dev.) The DZ11 driver is configured for two DZ11 (8 lines). The DC11 driver is set to run 4 lines. This can be changed in dc.c. The DH11 driver is set to handle 1 DH11 with a full complement of 16 lines. If you have less, or more, you may want to edit dh.c. The DN11 driver will handle 4 DN's. Edit dn.c. The DU11 driver can only handle a single DU. This can- not be easily changed. The KL/DL driver is set up to run a single DL11-A, -B, or -C (the console) and no DL11-E's. To change this, edit kl.c to have NKL11 reflect the total number of DL11-ABC's and NDL11 to reflect the number of DL11-E's. So far as the driver is concerned, the difference between the devices is their address. You should edit of the disk and tape drivers (rl.c, rf.c, rk.c, rp.c, tm.c, tc.c, ts.c, hk.c, hp.c, ht.c) to reflect the number of disk and tape drives in your configu- ration. The big disk drivers (hk.c, rp.c, and hp.c) have partition tables in them which you may want to experiment with. After all the corrections have been made, use `make(1)' to recompile the system (or recompile individually if you wish: use the makefile as a guide). If you compiled indi- vidually, say `make unix??' in the directory /sys/conf , the ?? is the CPU type, i.e., 23, 24, 34, 40, 60, 44, 45, 55, 70. The final object file will be named unix_ov for the PDP11/23, PDP11/24, PDP11/34, PDP11/40, and PDP11/60 or unix_id for the PDP11/44, PDP11/45, PDP11/55, and PDP11/70. This file should be moved to the root, and then booted to try it out. It is best to name it /nunix so as not to destroy the working system until you're sure it does work. See Boot Procedures(8) for a discussion of booting. Note: before taking the system down, always (!!) perform a sync(1m) to force delayed output to the disk. -14- SSppeecciiaall FFiilleess Next you must put in special files for the new devices in the directory /dev using mknod(1). Print the configura- tion file c.c created above. This is the major device switch of each device class (block and character). There is one line for each device configured in your system and a null line for place holding for those devices not config- ured. The essential block special files were installed above; for any new devices, the major device number is selected by counting the line number (from zero) of the device's entry in the block configuration table. Thus the first entry in the table bdevsw would be major device zero. This number is also printed in the table along the right margin. The minor device is the drive number, unit number or partition as described under each device in section 4 of the manual. For tapes where the unit is dial selectable, a spe- cial file may be made for each possible selection. You can also add entries for other disk drives. In reality, device names are arbitrary. It is usually convenient to have a system for deriving names, but it doesn't have to be the one presented above. Some further notes on minor device numbers. The hp and hk drivers use the 0100 bit of the minor device number to indicate whether or not to interleave a file system across more than one physical device. See hp(4) and hk(4) for more detail. The tm, ts, and ht drivers use the 0200 bit to indicate whether or not to rewind the tape when it is closed. The 0100 bit indicates the density of the tape, set for 800 BPI and cleared for 1600 BPI. By convention, tape special files with the 0200 bit on have an `n' prepended to their name, as in /dev/nmt0 or /dev/nrmt1. Again, see tm(4), ts(4) or ht(4). -15- The naming of character devices is similar to block devices. Here the names are even more arbitrary except that devices meant to be used for teletype access should (to avoid confusion, no other reason) be named /dev/ttyX, where X is some string (as in `00' or `library'). The files con- sole, mem, kmem, and null are already correctly configured. The makefile in /dev can be used to make the special files for the DZ11 and DH11 communications multiplexers as fol- lows: make tty_rm (removes all tty special files) make dz# (# = dz unit number) make dh# (# = dh unit number) Each dz11 unit handles 8 lines, and each dh11 unit 16 lines. Prior to making the dz or dh special files the "tty_rm" make command should be executed in order to remove an existing tty special files, the files "/dev/console" and "/dev/tty" will not be removed. The disk and magtape drivers provide a `raw' interface to the device which provides direct transmission between the user's core and the device and allows reading or writing large records. The raw device counts as a character device, and should have the name of the corresponding standard block special file with `r' prepended. (The `n' for no rewind tapes violates this rule.) Thus the raw magtape files would be called /dev/rmtX. These special files should be made. When all the special files have been created, care should be taken to change the access modes (chmod(1)) on these files to appropriate values (probably 600 or 644). -16- FFllooaattiinngg PPooiinntt UNIX only supports (and really expects to have) the FP11 type floating point unit. The PDP11/40 Floating Instruction Set is NOT supported by unix. For machines without the FP11 hardware, there is a user subroutine avail- able that will catch illegal instruction traps and interpret floating point operations. This allows programs with float- ing point constants to be compiled and executed, on machines without the FP11. To compile floating point programs use the `-f' flag to cc(1). This flag ensures that the floating point interpreter is loaded with the program and that the floating point version of `cc' is used. The system as delivered has this code included in only the iostat (1) and ac (1) commands and adb (1), although the operating system adapts automatically to the presence or absence of the FP11. If a program which has the floating point interpreter included is executed on a CPU with the FP11, the interpreter code is ignored and the floating point hardware is used instead. The floating point interpreter code may be removed by recompiling the program without the -f option. The floating point simulator software is extremely slow, if any serious floating point work is to be done the floating point hardware is really required. The changes described in the document "Unix Problems With Floating-Point Processors" by Bob Campbell, Ed Gould, Vance Vaughan, and Jim Reeds of the University of California at Berkeley, have been installed in this distribution. -17- TTiimmee CCoonnvveerrssiioonn If your machine is not in the Eastern time zone, you must edit (ed(1)) the file /sys/h/param.h to reflect your local time. The manifest `TIMEZONE' should be changed to reflect the time difference between local time and GMT in minutes. For EST, this is 5*60; for PST it would be 8*60. Finally, there is a `DSTFLAG' manifest; when it is 1 it causes the time to shift to Daylight Savings automatically between the last Sundays in April and October (or other algorithms in 1974 and 1975). Normally this will not have to be reset. When the needed changes are done, recompile and load the system using make(1) and install it. (As a general rule, when a system header file is changed, the entire system should be recompiled. As it happens, the only uses of these flags are in /sys/sys/sys4.c, so if this is all that was changed it alone needs to be recompiled.) You may also want to look at timezone(3) (/usr/src/libc/gen/timezone.c) to see if the name of your timezone is in its internal table. If needed, edit the changes in. After timezone.c has been edited it should be compiled and installed in its library. (See /usr/src/libc/(mklib and compall)) Then you should (at your leisure) recompile and reinstall all programs that use it (such as date(1)). -18- DDiisskk LLaayyoouutt The following table lists the file system interleave factors (m and n) to be used with mkfs (1m) and fsck -s (1m) to achieve optimum free list spacing when creating or sal- vaging file systems. These values were derived through experimentation, calculation and intuition, and although they may not be optimal in all cases they are certainly bet- ter that the default free list spacing assumed by mkfs and fsck. TABLE 1 File System Interleave Factors CPU RL01/2 RK06/7 RP03 RM02 RM03 RP04/5/6 TYPE m/n m/n m/n m/n m/n m/n 11/23 13/20 X X X X X 11/24 12/20 14/66 X 20/160 X 20/418 11/34 10/20 11/66 6/200 15/160 X 15/418 11/40 11/20 12/66 6/200 16/160 X 16/418 11/44 7/20 8/66 4/200 11/160 X 11/418 11/45 9/20 10/66 5/200 14/160 X 14/418 11/45z 11/20 12/66 X 17/160 X 17/418 11/55 9/20 10/66 5/200 14/160 X 14/418 11/60 9/20 10/66 5/200 14/160 X 14/418 11/70 6/20 6/66 X 9/160 13/160 9/418 If there are to be more file systems mounted than just the root ,/sys, and /usr, use mkfs(1) to create any new file system and put its mounting in the file /etc/rc (see init(8) and mount(1)). (You might look at /etc/rc anyway to see what has been provided for you.) There are two considerations in deciding how to adjust the arrangement of things on your disks: the most important is making sure there is adequate space for what is required; secondarily, throughput should be maximized. Swap space is a critical parameter. The system as distributed has 8778 (hpunix), 8778 (hkunix), 8000 (rpunix) , 3000 (rl01unix), 4000 (rl02unix) blocks for swap space. This should be large enough so running out of swap space never occurs. You may want to change these if local wisdom indicates otherwise. -19- Many common system programs (C, the editor, the assem- bler etc.) create intermediate files in the /tmp directory, so the file system where this is stored also should be made large enough to accommodate most high-water marks. If you leave the root file system as distributed (except as dis- cussed above) there should be no problem. All the programs that create files in /tmp take care to delete them, but most are not immune to events like being hung up upon, and can leave dregs. The directory should be examined every so often and the old files deleted. Exhaustion of user-file space is certain to occur now and then; the only mechanisms for controlling this phe- nomenon are occasional use of du(1), df(1), quot(1), threat- ening messages of the day, and personal letters. The efficiency with which UNIX is able to use the CPU is largely dictated by the configuration of disk con- trollers. For general time-sharing applications, the best strategy is to try to split user files, the root directory (including the /tmp directory) and the swap area among three controllers. Once you have decided how to make best use of your hardware, the question is how to initialize it. If you have the equipment, the best way to move a file system is to dump it (dump(1)) to magtape, use mkfs(1) to create the new file system, and restore (restor(1)) the tape. If for some rea- son you don't want to use magtape, dump accepts an argument telling where to put the dump; you might use another disk. Sometimes a file system has to be increased in logical size without copying. The super-block of the device has a word giving the highest address which can be allocated. For rel- atively small increases, this word can be patched using the debugger (adb(1)) and the free list reconstructed using icheck(1). The size should not be increased very greatly by this technique, however, since although the allocatable space will increase the maximum number of files will not (that is, the i-list size can't be changed). Read and understand the description given in file system(5) before playing around in this way. You may want to see section rp(4) or hp(4) for some suggestions on how to lay out the information on RP disks. More detailed information about the disk partitions may be obtained from the `sizes' tables in the disk drivers (hp.c, hk.c, rp.c). Also see /usr/doc/hksizes and /usr/doc/hpsizes for the RK06/7 and RM02/3, RP04/5/6 disk layouts. -20- If you have to merge a file system into another, exist- ing one, the best bet is to use tar(1). If you must shrink a file system, the best bet is to dump the original and restor it onto the new filesystem. However, this might not work if the i-list on the smaller filesystem is smaller than the maximum allocated inode on the larger. If this is the case, reconstruct the filesystem from scratch on another filesystem (perhaps using tar(1)) and then dump it. If you are playing with the root file system and only have one drive the procedure is more complicated. What you do is the following: 1. GET A SECOND PACK!!!! 2. Dump the current root filesystem (or the reconstructed one) using dump(1). 3. Bring the system down and mount the new pack. 4. Retrieve the WECo distribution tape and perform steps 1 through 5 at the beginning of this document, substitut- ing the desired file system size instead of 9600 when asked for `file system size'. 5. Perform step 6 above up to the point where the `tape' question is asked. At this point mount the tape you made just a few minutes ago. Continue with step 6 above substituting a 0 (zero) for the 5. -21- NNeeww UUsseerrss Install new users by editing the password file /etc/passwd (passwd(5)). This procedure should be done once multi-user mode is entered (see init(8)). You'll have to make a current directory for each new user and change its owner to the newly installed name. Login as each user to make sure the password file is correctly edited. For exam- ple: ed /etc/passwd $a joe::10:1::/usr/joe: w q mkdir /usr/joe chown joe /usr/joe login joe ls -la login root This will make a new login entry for joe, who should be encouraged to use passwd(1) to give himself a password. His default current directory is /usr/joe which has been cre- ated. The delivered password file has the user _b_i_n in it to be used as a prototype. -22- MMuullttiippllee UUsseerrss If UNIX is to support simultaneous access from more than just the console terminal, the file /etc/ttys (ttys(5)) has to be edited. To add a new terminal be sure the device is configured and the special file exists, then set the first character of the appropriate line of /etc/ttys to 1 (or add a new line). If the new terminal is to be a "local" terminal, i.e., not a dialup line, set the first character to 2, this will enable the terminal to operate without hav- ing carriar asserted. Note that init.c will have to be recompiled if there are to be more than 100 terminals. Also note that if the special file is inaccessible when init tries to create a process for it, the system will thrash trying and retrying to open it. FFiillee SSyysstteemm HHeeaalltthh Periodically (say every day or so) and always after a crash, you should check all the file systems for consistency (fsck (1m)). You should create the file "/etc/checklist" containing the names of the file systems to be checked, see fsck (1m) for more detail. It is quite important to execute sync (1m) before rebooting or taking the machine down. This is done automatically every 30 seconds by the update program (8) when a multiple-user system is running, but you should do it anyway to make sure. Dumping of the file system should be done regularly, since once the system is going it is very easy to become complacent. Complete and incremental dumps are easily done with dump(1). Dumping of files by name is best done by tar(1) but the number of files is somewhat limited. Finally if there are enough drives entire disks can be copied using cp(1), or preferably with dd(1) using the raw special files and an appropriate block size. -23- CCoonnvveerrttiinngg SSiixxtthh EEddiittiioonn FFiilleessyysstteemmss The best way to convert file systems from 6th edition (V6) to 7th edition (V7) format is to use tar(1). However, a special version of tar must be prepared to run on V6. The following steps will do this: 1. change directories to /usr/src/cmd/tar 2. At the shell prompt respond make v6tar or make v6tar40 This will leave an executable binary named `v6tar' or `v6tar40'. 3. Mount a scratch tape. 4. Use tp(1) to put `v6tar' on the scratch tape. 5. Bring down V7 and bring up V6. 6. Use tp (on V6) to read in `v6tar'. Put it in /bin or /usr/bin (or perhaps some other preferred location). 7. Use v6tar to make tapes of all that you wish to con- vert. You may want to read the manual section on tar(1) to see whether you want to use blocking or not. Try to avoid using full pathnames when making the tapes. This will simplify moving the hierarchy to some other place on V7 if desired. For example chdir /usr/ken v6tar c . is preferable to v6tar c /usr/ken 8. After all of the desired tapes are made, bring down V6 and reboot V7. Use tar(1) to read in the tapes just made. -24- OOddddss aanndd EEnnddss The programs dump, quot, ncheck, and df (source in /usr/source/cmd) should be changed to reflect your default mounted file system devices. Print the first few lines of these programs and the changes will be obvious. Tar should be changed to reflect your desired default tape drive. Good Luck Charles B. Haley Dennis M. Ritchie Fred Canter