PAGESWAPPER Volume Shadowing on VMS Massbus Disks by Allen A. Watson, Manager Systems Technology The Record (newspaper) 150 River Street, Hackensack, NJ 07602 Dear Larry: I just received the September Pageswapper, the one with only eight pages, and decided to take pity and send you what little I have as a new VAX user before my duties as new co-editor of the RSX Multitasker give me an excuse not to write. We are RSX11M-Plus users migrating to VMS on VAX 11/780's. One of the main reasons we purchased M-Plus was its capability of volume shadowing -- constant on-line mirror-image copying of a disk onto another for total backup. When we began to consider moving to the VAX, last January, we were dismayed to learn VMS did not support shadowing. We made having shadowing a condition of sale for two 780's, and DEC found a way to do it. A VMS shadowing package was developed in the DEC office in Munich, Germany. We purchased it from DEC, through our local sales office in Piscataway, New Jersey. It works very much like M-Plus shadowing and supports Massbus disks. We have ten 300 Mb disks -- they are not Digital disks or controllers -- and VMS shadow does work. 1 PAGESWAPPER - December 1983 - Volume 5 Number 6 Volume Shadowing on VMS Massbus Disks VMS Shadown requires either that all disks in a shadow set are error-free (zero bad blocks), or that you use BAD to perform a logical OR of the Bad Block files so that all disks have the same blocks marked bad. In Las Vegas, I asked a DEC speaker why the same concept could not work for Massbus disks under Version 4.0. Digital has indicated support of shadowing only through the HSC-50 controller, or only for the new RA series disks. The speaker stated there was no reason it could not be done unless the MSCP Server task was written to specifically exclude it. He stated that exclusion had not been done, and he had no intention of putting it in. Later in the same meeting, a Digital employee from the Washington DC office informed me that a project involving shadowing of Massbus disks on a VAX Cluster had already been completed for the MCI telephone network. I do not feel at liberty to give out the names of Digital employees in these offices. That would not be fair to these individuals who have gone out of their way to meet the needs of their customers. I suggest that readers who are interested in volume shadowing for Massbus disks contact their local sales representatives armed with the knowledge that it does exist and has been sold to __ ____ _____ VAX users by Digital. Let them (your local rep) make the __ _______ necessary contacts in the offices I have named above. Secondly, again for users of Massbus disks, I want to pass along the information that you do not have to spend $100,000 to by VAX cluster hardware to enter the world of clusters. The implications have been that a minimal cluster must include an _______ HSC-50 and RA-type disk -- not so. A Star coupler and a dual CI are all you need to get many of the ____ benefits of the new cluster software. Specifically, you get: - Total sharing of all disks among all CPU's in the ___ cluster -- data is globally available. - Cluster-wide record locking, enabling multiple file access. - Disk performance optimization through the MSCP server task that approaches the performance of disks on the HSC-50. - Sharing of other peripherals such as printers. If you want to start growing into a cluster with a limited budget, this is the way to go. 2 PAGESWAPPER - December 1983 - Volume 5 Number 6 In this issue... In this issue... Volume Shadowing on VMS Massbus Disks . . . . . . . 1 In this issue... . . . . . . . . . . . . . . . . . . 3 Editor's Workfile . . . . . . . . . . . . . . . . . 4 The System Improvement Request Program Scorecard . . 5 VAX Bus Overview . . . . . . . . . . . . . . . . . 10 VMS/Network Working Group . . . . . . . . . . . . 18 Using SDA and CCL . . . . . . . . . . . . . . . . 19 INPUT/OUTPUT . . . . . . . . . . . . . . . . . . . 43 INPUT/OUTPUT Submission Form . . . . . . . . . . . 50 System Improvement Request Submission Form . . . . 52 General material for publication in the Pageswapper should be sent (US mail only -- no "express" services please) to: Larry Kilgallen, PAGESWAPPER Editor Box 81, MIT Station Cambridge, MA 02139-0901 Preference is given to material submitted as machine-readable text (best is Runoff source). Please do not submit program source, as that is better distributed on the VAX SIG tape. Material for "The DBMS Monitor" section of the Pageswapper (pertaining to VAX-11 DBMS) should be sent to: Julie Llewellyn United Technologies Microelectronics Center 1365 Garden of the Gods Road Colorado Springs, CO 80907 Change of address, reports of non-receipt, and other circulation correspondence should be sent to: DECUS U.S. Chapter, MRO2-1/C11 Attention: Publications Department One Iron Way Marlborough, MA 01752 USA Only if discrepancies of the mailing system are reported can they be analyzed and corrected. 3 PAGESWAPPER - December 1983 - Volume 5 Number 6 Editor's Workfile Editor's Workfile by Larry Kilgallen, Pageswapper Editor If you missed the Las Vegas session on VAX buses, be sure to see the article in this issue. Pay particular attention to the material on the General Purpose Backplane Bus, which DEC has not announced but is obviously working on. About three days after the November issue of the Pageswapper went to production, I had the opportunity to observe the DEC stockholders meeting (speaking as a resident of Boston's Back Bay section, I must compliment DEC on their choice of location for that meeting and for DECtown). In the equipment displayed to impress the stockholders were not only the MicroVAXs shown at the Las Vegas DECUS symposium (I wonder if hardware can qualify for the airlines' frequent flyer plans?), but also a snappy new graphic display attached to one of them. The display had various lines and circles and other graphic-type stuff on it (that's a different newsletter, folks) and also multiple simultaneous VT100 emulations (accomplished by interfacing as a terminal port driver, according to the developers who were there). The windows for the various emulated VT100's were only above and below one-another, no side-by-side. The most amazing thing about the display terminal, however, was the SPEED with which it switched back and forth between two pictures (one doing graphics and one being a set of VT100's in this case). The switching (I was told) is done with memory mapping registers (virtual mirrors?). The display is implemented with shared memory, taking up 256 Kbytes of Qbus address space and in this case using 128 Kbytes at a time. The instantaneous change I was seeing was the switch back and forth between the two pre-loaded areas of memory. So if you have a potential MicroVAX application where you have some advance idea of what the next frame is going to look like, you can join me in hoping this device will be marketed. There may be some interest in higher places though, the president of DEC told the audience of assembled stockholders that the MicroVAX/bit-mapped terminal combination was "the personal computer of the not-too-distant future. It is the one I want and the one most people will want." Other quotations from chairman Ken: o "In the next few months, when we announce our second MicroVAX chip..." o On the future high-speed VAX being late, he said he had "no apologies for what we're doing here" and that the machine was "by all definitions a mainframe computer". 4 PAGESWAPPER - December 1983 - Volume 5 Number 6 Editor's Workfile o On DEC involvement with (and investment in) Trilogy technology he said he was "willing to gamble 26 million dollars on it, but not willing to gamble the future of the VAX on it". The System Improvement Request Program Scorecard Gary L. Grebus System Improvement Request Coordinator THE HISTORY We all have ideas about how the VAX hardware or software could be improved. The trick has always been to communicate these good ideas to Digital. To help with this problem, the VAX Systems SIG created the Systems Improvement Request (SIR) program in late 1979. Since then, there has been a steady stream of improvement requests from the VAX user community. The SIG has conducted five formal ballots to rank these requests, and DEC has responded to the top ten items from each ballot. THE RESULTS The SIR program has always drawn many positive comments. It is valued by the SIG and by DEC as an important communication channel. But the real question is whether user needs, as represented by the top SIR's, are being communicated to DEC. To try to measure our success, I have collected the top ten items from all the ballots, divided into (perhaps slightly subjective) categories of "Done", "Won't Be Done", and "Not Done". The first category includes items currently available on VAX systems, and those expected to appear with Version 4.0 of VMS. The second category contains those items where the DEC response indicated that they would not implement the request. The justifications for these responses is included. The last category groups those items which DEC has indicated as desirable but which have so far escaped implementation. This category is somewhat inflated in that four of its items represent new requests within the last six months. The results of this tally are: Done 39 (58 %) Won't Be Done 9 (18 %) Not Done 12 (24 %) Based on these statistics, it seems clear that there is fairly 5 PAGESWAPPER - December 1983 - Volume 5 Number 6 The System Improvement Request Program Scorecard good communication of those user concerns represented through the SIR process. DEC has responded positively to a large majority of the top ten requests. Restricting consideration to only top ten items certainly under estimates the benefit of the SIR program. A large number of requests which have not made the top ten have still become part of VMS. HOW TO WRITE A SYSTEM IMPROVEMENT REQUEST The key to the success of the SIR process is good communications. This process begins with SIR's you submit. The more clearly you describe your needs, the better the SIR ballot will become. Here are some guidelines to consider when writing a request: o Make your request in terms of the capabilities required. This prevents the issue from being cluttered by the pros and cons of a suggested implementation. For example, a request to "reduce the system I/O overhead involved in paging" is more useful than one to "increase the size of a VAX page to 4096 bytes". If you wish to suggest an implementation, do so after describing the requirements. o Provide as much detail as possible about the capability you are requesting. Make sure the expected advantages are clear. Use examples where possible. Don't just reference a feature in another operating system; no one may be familiar with it. o If something is really a bug, send an SPR to Digital. The SIR process is oriented toward features enhancements and operates on a fairly slow time scale. Collected Top Ten SIR's SIR's Which Have Been Implemented: 1. Date-last-accessed for files 2. Enhanced performance monitor with hardcopy 3. NAMELIST I/O in FORTRAN 4. SET PROCESS/SUSPEND and RESUME 5. User extensible CLI 6. Multiprocessing capability for each user (SPAWN) 7. Checkpoint/restart for some jobs (V4) 6 PAGESWAPPER - December 1983 - Volume 5 Number 6 The System Improvement Request Program Scorecard 8. Display program input in batch logs (V4) 9. Add terminal device name to accounting record 10. Cross-reference listing for FORTRAN 11. Control of batch log file disposition 12. More description of system tuning (twice in top ten) 13. Disk verifier should understand ACP caching 14. Default file protection to that of an older version (V4) 15. Don't kill process when phone line lost (V4) 16. Access control lists for files (V4) 17. Command line edit and reissue (V4) 18. Better control of operator messages 19. Directory by UIC (V4) 20. COPY/OWNER qualifier (V4) 21. Application oriented manual on System Services (V4) 22. Multiple level of control for BROADCAST messages (V4) 23. Enhanced editing of user input to DCL procedures (V4) 24. Reporting of bad block info by file name (V4) 25. Extensions to rooted directory syntax (V4) 26. Image accounting on selected images (V4) 27. More flexible control over disk quota (partially done in V4) 28. Set file modify date on COPY (V4) SIR's Which Won't Be Implemented 1. Auto-logout of inactive terminals. (Can be done with simple programs. Too dependent on site policies) 2. Default ownership of created file to directory owner. (Conflicts with VMS schemes of resource accountability) 7 PAGESWAPPER - December 1983 - Volume 5 Number 6 The System Improvement Request Program Scorecard 3. Provide hooks for LOGOUT extensions or command procedures. (Only a very restricted context can be guaranteed at process deletion time). 4. Use Login/Logout disk quotas (This scheme has too many loopholes.) 5. General foreign terminal support (Done for VMS utilities but not for EDT and layered products.) 6. Provide a tape transfer utility (VMS processes ANSI standard tapes.) 7. Provide a SPAWN/CONTINUE qualifier (This does not make sense in the general case involving ATTACH command) 8. Provide variable tab stops in EDT 9. Add "unprintable data" flag to file header (This problem would be better addressed with a filter in the print symbiont) SIR's Not Yet Implemented 1. Enhanced privilege for command procedures. Twice in the top ten. (No truly secure approach has been found) 2. Execute-only command procedures 3. Report which quota has been exceeded. (This has been done for new quotas. There are serious problems of upward compatability) 4. Operator initiated rollout of processes. Twice in the top ten. (This is a very complex problem which has yet to be solved. It is complicated by VAX-clusters) 5. Ability to log CRT activity to a file. 6. Structured programming features for DCL. 7. Full support for batch and print queues over network. (Problem must first be addressed by DECnet architecture) 8. $GETJPI should be able to obtain default directory and device. 9. Support DCL tables in source form. (Currently, too many undocumented CLI features are used in the tables) 8 PAGESWAPPER - December 1983 - Volume 5 Number 6 The System Improvement Request Program Scorecard 10. Support return-receipt for MAIL messages. 9 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview VAX Bus Overview Dileep P. Bhandarkar VAX Systems Architecture, DEC Editor's Note The following material is taken from slides used in presenting the subject at the Fall 1983 (Las Vegas) DECUS Symposium. Levels of Interconnection o Local Area Network - Heterogenous systems - Distances of a few kilometers - Large numbers of nodes (100's) o Clusters - Closer coupling between systems - Typically within one room - Smaller number of nodes (10's) o Intra-system - I/O adapters - CPU to memory connection - Short length 10 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview Local Area Networks o Network interconnect -- ETHERNET o 10 Megabits/second bandwidth o Up to 1.5 kilometers distance with 2 repeaters o Reliable multi-drop topology o Large number of drops o Industry standard Ethernet Usage o Interconnect for DECnet nodes in a building o Interconnect for shared servers - Terminal servers - Print servers - File servers - Gateways to other networks 11 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview Clusters o Computer interconnect o 70 Mbits per second o Logically multi-drop o Physically a star configuration with passive hub o Up to 45 meters from star coupler to node o Up to 16 nodes o Dual path o Separate transmit and receive coaxial cables o packet oriented Computer Interconnect Usage o Connection to intelligent storage controllers o Mechanism for incremental growth of compute power o Basis for high availability system implementation 12 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview Intra-system Buses o Internal bus - SBI on VAX-11/780 - CMI on VAX-11/750 - Unique to each CPU o External bus - QBUS - UNIBUS - Unique for a family of products QBUS Characteristics o 22-bit address space o 44 signal lines - multiplexed data/address lines o Single bus arbiter o Asynchronous operation o Block mode DMA transfer capability o Allows multi-backplane configuration with 16 foot cables o Typical bandwidth between 1 and 3 MB/second o Easy to interface 13 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview Unibus Characteristics o 18 bit addressing o 56 signal lines o Separate address and data lines o 4 interrupt priority levels o Single bus arbiter o Asynchronous operation o Allows multiple backplane configurations with 50 foot cable o Typical bandwidth between 1 and 2 MB/second Intra-System Internal Bus o Allows attachment of adapters to different external busses o VAX-11/780 SBI attachments - MASSBUS - UNIBUS - CI - DR32 - memory - multi-port memory o Adapters to internal bus must be re-engineered for each CPU 14 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Overview SBI characteristics o 84 signal lines o Length less than 10 feet o 30-bit address space o Synchronous operation with 200 ns clock o Distributed arbitration o Parity o 4 interrupt levels o Pended bus o Up to 16 bus attachments o 13.3 MB/sec bandwidth Intra-System buses o Inappropriate to standardize on single bus - Wide range of cost and performance - Different technologies o Standard neede for external bus - UNIBUS bandwidth insufficient - UNIBUS based on 15 year old technology - No current industry standard for 32-bit bus 15 PAGESWAPPER - December 1983 - Volume 5 Number 6 General Purpose Backplane Bus General Purpose Backplane Bus Desired characteristics o 30-bit address capability o High bandwidth -- greater than 10 megabytes/second o Processor-memory and I/O interconnect o Reliability features o Standard interface with set of LSI chips o Must support requirements of VAX architecture o Multiprocessor support Anticipated usage o Similar to role of UNIBUS on PDP-11 systems o System bus for small systems o I/O bus for large systems o Primary bus for general purpose I/O interfaces - terminal multiplexers - disk/tape controllers - communications interfaces o UNIBUS adapter for migration of existing controllers 16 PAGESWAPPER - December 1983 - Volume 5 Number 6 VAX Bus Strategy VAX Bus Strategy o ETHERNET for local area networks and shared servers o CI based clusters for mid-range and large processors o Current VAX processors support UNIBUS o MicroVAX provides VAX functionality on QBUS o Migration to new 32-bit bus for general-purpose I/O Regarding the future of the MASSBUS, Mr. Bhandarkar said: o No new Massbus disks are planned o No new Massbus adapters are being developed o New processors may support the Massbus if they contain an SBI or CMI 17 PAGESWAPPER - December 1983 - Volume 5 Number 6 VMS/Network Working Group VMS/Network Working Group by Todd L. Voros Room 168, Wisconsin Telephone Company 845 North 35th Street, Milwaukee, WI 53208 This group has simultaneous existence in two SIGs: the VAX Systems SIG and the Networks SIG. It represents VMS interests to the Networks SIG and network interests to the VAX Systems SIG. Twelve people attended the working group meeting at the Fall 1983 (Las Vegas) DECUS Symposium. Surveying the attendees: 1. 7 have DECnet 2. 7 have Multinet 3. 5 have Network Control organizations 4. 4 are planning to install DECnet 5. 10 have Local Area Networks: 1. 7 baseband 2. 3 broadband 2 have large IBM shops An interest was expressed in holding sessions at the next DECUS Symposium (Cincinnati, OH) dealing with specific networking solutions developed by group members. The major areas of concern for the group appear to be: 1. Local Area Networks - Problems and their resolutions 2. Ability to network different vendors equipment together (IBM to DEC, etc.) 3. User-written DECnet drivers are desired by some group members 18 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL Using SDA and CCL INVOCATION CCL is invoked from the console terminal with ^P. The CCL prompt is >>> SDA is invoked from dcl with the ANALYZE/SYSTEM or the ANALYZE/CRASH commands. the SDA prompt is SDA> the first is used to analyze the running system. The other is used to analyze a crash dump. CCL is only used on the running system (when halted). commands To use most CCL commands the CPU must be halted. This is accomplished using the HALT instruction (abbreviated H). To unhalt the CPU, the CONTINUE command is used (abbreviated C). CCL examine the CCL examine command has the form: EXAMINE (abbreviated E) with qualifiers and an address. EXAMINE/VIRTUAL treats the address argument as a virtual address. EXAMINE/BYTE,EXAMINE_/WORD,EXAMINE_/LONG specify length. EXAMINE/GENERAL is used to examine registers. EXAMINE/NEXT:n examines n locations. EXAMINE/INTERNAL examines a processor register. The other qualifiers available with the EXAMINE command are /quad,/physical, /v bus which are not discussed here. 19 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL CCL DEPOSIT The CCL DEPOSIT COMMAND has the form: DEPOSIT qualifiers address data the qualifiers for the CCL DEPOSIT command are the same as for the examine command CCL SET the CCL SET COMMAND has the following forms: SET DEFAULT default options where the options are the same as the qualifiers for the EXAMINE command. SET STEP SET TERMINAL PROGRAM SET SOMM SET CLOCK (SLOW or FAST or NORMAL) SET RELOCATION OTHER CCL COMMANDS the TEST command (abbreviated T) is used without qualifiers to invoke the microdiagnostic tests. If no errors are detected control is returned to CCL. Other commands are BOOT,CLEAR,HELP (abbreviated HE),INITIALIZE,LINK,LOAD, PERFORM,QUAD,CLEAR,REBOOT,REPEAT,SHOW,START,NEXT,UNJAM,WAIT, and WCS. SDA EXAMINE the SDA EXAMINE command, unlike the CCL EXAMINE, only examines virtual addresses and will not examine hardware registers. The SDA EXAMINE however will translate symbolic addresses and SDA has commands which will format structures. Allowing one to obtain various information, without following complex sets of pointers. The form of the examine command is: EXAMINE qualifiers range A range is either a location , a location:location or a location;length. 20 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL There are five qualifiers available with the examine command: /p0 /p1 /system /all /instruction The first four of the qualifers are used to display large portions of core. EXAMINE/INSTRUCTION is used to display a range as instructions. SDA EXIT Exit is used to get out of SDA. SDA FORMAT FORMAT is used to display an area according to some symbolic definitions that are either known to SDA, or are read in. SDA SET There are two SDA set commands. the first SET OUTPUT file-spec is used to have the output from SDA go to a file rather than the terminal. the other SET PROCESS is used to set the default process (originally the current process) for SDA SHOW commands and for such commands as EXAMINE/P0. it has three forms: SET PROCESS name SET PROCESS/INDEX=n SET PROCESS/SYSTEM where n is the right half of the processes PID, and where /SYSTEM causes information about the system to be displayed. 21 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL SDA SHOW There are nine SDA show commands they are: SHOW CRASH SHOW DEVICE SHOW PAGE_TABLE SHOW PFN_DATA SHOW POOL SHOW PROCESS SHOW STACK SHOW SUMMARY SHOW SYMBOL The command SHOW PROCESS is used to display information about a specific process. It has the form SHOW PROCESS parameter qualifiers The parameters are: name /index=n /system These are the same as in the SET PROCESS and determine for which process information will be displayed. If they are omitted the default will be taken (see SET PROCESS). The qualifiers are: /PCB the default /PHD /REGISTERS /WORKING_SET /PROCESS_SECTION_TABLE /PAGE_TABLES /ALL These are used to control which information is displayed. DISPLAYING INFORMATION 22 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL displaying information about a process from the PCB one can obtain information about a process by display data from the process PCB. to do this in SDA one would use the command: SHOW PROCESS in CCL one would have to obtain this information through a series or examines. the results of an SDA show process command are shown below. Process 43 dump: SYSPG1 Process status: 00040001 RES,PHDRES PCB address 800F1D40 JIB address 800F5940 Master PID 00110043 Creator PID 00000000 PID 00110043 Subprocess count 0 PHD address 803EAE00 Swapfile disk address 01002681 State CUR Termination mailbox 0000 Current priority 8 AST's enabled KESU Base priority 4 AST's active NONE UIC [002,002] AST's remaining 38 Mutex count 0 Buffered I/O count/limit 6/6 Waiting EF cluster 0 Direct I/O count/limit 6/6 Starting wait time 00001B1A BUFIO byte count/limit 3936/4096 Event flag wait mask F7FFFFFF # open files allowed left 15 Local EF cluster 0 CC00001F Timer entries allowed left 10 Local EF cluster 1 80000000 Active page table count 0 Global cluster 2 pointer 00000000 Process WS page count 436 Global cluster 3 pointer 00000000 Global WS page count 28 WHAT DO TO WHEN SDA REFUSES TO DISPLAY THIS INFORMATION In certain cases (such as when a process is in MWAIT) SDA will say that it is unable to access some virtual address and will not do a show process. One can obtain the same information, although not as easily, in the following manner. First one does SHOW SYMBOL SCH$GL_PCBVEC One takes the value stored at this location and adds four times the process index. one then takes the result of this calculation n and does EXAMINE n one then takes the value m obtained and does a FORMAT m This will also display the information from the process PCB. In fact more information is displayed this way then with the SHOW PROCESS command. 23 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL PROCESS NAME AND INDEX The first line of this display shows the process name and the process index (which is the right half of the PID). PROCESS STATUS next the process status flags are shown. In the example, the process is resident in core and its header is resident. The various flags are: ASTPEN AST is pending BATCH DELPEN delete is pending DISAWS disable automatic working set adjustment FORCPEN forced exit is pending HIBER process should hibernate after image activation INQUAN process is in its initial quantum (following inswap) LOGIN process should be allowed to login without reading SYSUAF NETWRK process is a network job NOACNT do not write an accounting record for this process NODELET do not delete this process PHDRES process header is resident PSWAPM process swapping is disabled PWRAST process has declared a power recovery AST RES process is resident RESPEN resume is pending SSFEXC enable system service exceptions for kernel mode SSFEXCE for executive mode SSFEXCS for supervisor mode SSFEXCU for user mode SSRWAIT disable resource wait mode SUSPEN suspend is pending SWPVBN modified page write to the swap file is in progress WAKEPEN wake is pending WALL wait for all event flags in mask These flags are all defined as PCB$V_flagname both in SYS.MAP and by the $PCBDEF macro. for further information on these flags see the SYSTEM SERVICES MANUAL: $CREPRC for BATCH,HIBER,LOGIN,NETWRK,NOACNT,PSWAPM,SSFEXCU,SSRWAIT. $SETPRA for PWRAST. $SETRWM for SSRWAIT. $SETSWM for PSWAPM. $SETSFM for SSFEXC,SSFEXCE,SSFEXCS,SSFEXCU. $WFLAND for WALL. Also see DCL MANUAL: RUN for HIBER,LOGIN,PSWAPM,SSFEXCU. 24 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL PCB ADDRESS AND JIB ADDRESS The next line contains the PCB address and the JIB address. These addresses can be used to reference various parts of the PCB and JIB. MASTER PID AND CREATOR PID The next line gives the master PID and the creator PID. for a subprocess the creator PID is the PID that created it and the master PID is the PID at the top of the tree of subprocess creations. PID AND SUBPROCESS COUNT The next line gives the PID and the subprocess count. the PID is the major index of the process leading to pointers to all the other process information. The important part of the PID is the right hand part called the INDEX. the left hand part of the PID is simply a sequential count (reinitialized on boot) and incremented each time a new process is given this particular index. the subprocess count is mainly used by the system when deleting a process. PHD ADDRESS AND SWAPFILE DISK ADDRESS The next line gives the PHD address and the swapfile disk address. The PHD (process header) address can be used as a pointer into the process header to obtain information contained therein. The swapfile disk address is used by the swapper for in and outswaps. 25 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL STATE AND TERMINATION MAILBOX The next line gives the state and the termination mailbox number. the states are 1 COLPG collided page wait 2 MWAIT mutex or misc. resource wait 3 CEF common event flag wait 4 PFW page fault wait 5 LEF local event flag wait 6 LEFO local event flag wait outswapped 7 HIB hibernate 8 HIBO hibernate outswapped 9 SUSP suspended A SUSPO suspended outswapped B FPG free page wait C COM compute D COMO compute outswapped E CUR current Certain process send a mailbox message upon completion (this is used for instance in the notify on the SUBMIT command). If the process does this then the termination mailbox number will contain the number of this mailbox, otherwise it will be zero. CURRENT PRIORITY AND ASTS ENABLED The next line has the current priority (which is the scheduling priority) and the ASTS enabled. ASTS can be enable or disable for each of the access modes Kernel,Executive, Supervisor, and User. The letters listed are the first letters of the mode names for which AST may be received. BASE PRIORITY AND AST ACTIVE The next line contains the base priority and the access modes for which ASTs are active. The base priority is the priority to which a process drops after recieving a compute quantum various events such as i/o completion raise the current priority from the base priority. When a process recieves an AST it can still receive ASTs for lower (more priveleged) access modes; therefore there may be up to four ASTs active one for each of Kernel,Executive,Supervisor, and User. 26 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL UIC AND ASTS REMAINING The next line contains the user identification code and the number of ASTs remaining. Every process has an ASTLM which is the number of AST which it may have outstanding when the process causes an AST to be declared one is subtracted from that count and when the AST is delivered one is added to that count. This value is the resulting count. MUTEX COUNT AND BUFFERED I/O COUNT/LIMIT The mutex count is the count of the number of system data structures on which the process has a lock. The buffered i/o count is the number of outstanding buffered i/o requests the process has and the buffered i/o limit is the number the process is authorize to have (biolm). WAITING EF CLUSTER AND DIRECT I/O COUNT/LIMIT If a process is in LEF the waiting event flag cluster tells which cluster the flag (or flags depending on whether wall is set) is in that the process is waitng for. The direct i/o count is the number of outstanding direct i/o requests the process has and the direct i/o limit is the number the process is authorize to have (diolm). STARTING WAIT TIME AND BUFIO BYTE COUNT/LIMIT The bufio count is the number of bytes of system memory being used as buffers for this process. The limit is the authorized amount the process is allowed to have (bytlm). 27 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL EVENT FLAG WAIT MASK AND # OPEN FILES ALLOWED LEFT The event flag wait mask has two uses: 1 if the process is in LEF or CEF the event flag wait mask has a 0 for each flag the process is waiting for. 2 if the process is in MWAIT then the event flag wait mask has a reason code for the MWAIT. there are two types of reason codes those which start with an 8 (i.e. are system addresses) which must be looked up in SYS.MAP and those which are small numbers these are: 1 ast wait (wait for system or special kernel mode AST 2 mailbox full 3 nonpaged dynamic memory 4 page file full 5 paged dynamic memory 6 breakthrough (wait for broadcast message) 7 image activation lock 8 job pooled quota 9 lockid 10 swapfile 11 modified page list empty 12 modified page writer busy The number of open files allowed left is the authorized number of open files (fillm) minus the number of files open. LOCAL EVENT FLAG CLUSTER 0 AND TIMER ENTRIES ALLOWED LEFT the local event flag cluster 0 is simply the value of event flags 0 thru 31. The timer entries allowed left is the authorize number of timer queue entries (tqelm) minus the number of timer queue entries (requests to be notified or have an event flag set in a certain amount of time) that the process has outstanding. LOCAL EVENT FLAG CLUSTER 1 AND ACTIVE PAGE TABLE COUNT The local event flag cluster 2 is simply the value of event flags 32 thru 63. 28 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL GLOBAL CLUSTER 2 POINTER AND PROCESS WS PAGE COUNT The global cluster 2 pointer is a pointer to a common event flag block which contains flags 64 thru 95. a common event flag block has the form 0 forward link 4 backward link 8 size A type B status C creator PID 10 event flags 14 waiting queue forward link 18 waiting queue back link ... 20 uic of creator 28 cluster name as a counted ascii string The process WS page count is the number of pages in the working set GLOBAL CLUSTER 3 POINTER AND GLOBAL WS PAGE COUNT The global cluster 3 pointer is a pointer to the common event flag block for flags 96 thru 127. The global ws page is count is the number of global pages in the working set. display information from the ceb As an example if one wished to display the flags for cluster 2 one would take the value in the pcb for global cluster 2 add 10 to that and do an E that number 29 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL DISPLAYING THE ABOVE INFORMATION USING CCL First one must obtain the PCB address. For the current process this is located in SCH$GL_CURPCB. looking this value up in SYS.MAP on obtains a value (for the current version of the operating system). The value of SCH$GL_CURPCB for version 3 is 8000210C. One then does an E/V 8000210C (or the appropriate value if not under V3). Then one goes back to SYS.MAP and looks up the PCB location one is interested in adds that number to the value you obtained from the examine and does either an E/V or an E/V/B or and E/V/W depending on the size of the value one is interested in. For instance, with the current process being as it was in the SHOW PROCESS at the beginning of this chapter if one wanted to display the Event flag wait mask and the state for the current process one would do the following. E/V 8000210C the machine would respond with the physical value of 8000210C and 800F1D40 then one would look up state and event flag wait mask in SYS.MAP and find that PCB$W_STATE = 2C and PCB$L_EFWM = 4C then one would add 800F1D40 + 2C = 800F1D6C and 800F1D40 + 4C = 800F1D8C one would E/V/W 800F1D6C and get 000E which means current and E/V 800F1D8C and get F7FFFFFF which is the event flag wait mask here are the current PCB values from SYS.MAP PCB$B_ASTACT 0000000C PCB$B_ASTEN 0000000D PCB$B_PRI 0000000B PCB$B_PRIB 0000002F PCB$B_PRIBSAV 00000029 PCB$B_PRISAV 00000028 PCB$B_TYPE 0000000A PCB$B_WEFC 0000002E PCB$C_LENGTH 0000009C PCB$K_LENGTH 0000009C PCB$L_ARB 00000084 PCB$L_ASTQBL 00000014 PCB$L_ASTQFL 00000010 PCB$L_DLCKPRI 00000094 PCB$L_EFC2P 00000058 PCB$L_EFC3P 0000005C PCB$L_EFCS 00000050 PCB$L_EFCU 00000054 PCB$L_EFWM 0000004C 30 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL PCB$L_IPAST 00000098 PCB$L_JIB 00000078 PCB$L_LOCKQBL 00000090 PCB$L_LOCKQFL 0000008C PCB$L_OWNER 0000001C PCB$L_PHD 00000064 PCB$L_PHYPCB 00000018 PCB$L_PID 00000060 PCB$L_PQB 0000004C PCB$L_SQBL 00000004 PCB$L_SQFL 00000000 PCB$L_STS 00000024 PCB$L_UIC 00000088 PCB$L_WSSWP 00000020 PCB$L_WTIME 00000028 PCB$Q_PRIV 0000007C PCB$S_LNAME 00000010 PCB$S_PRIV 00000008 PCB$S_TERMINAL 00000008 PCB$T_LNAME 00000068 PCB$T_TERMINAL 00000044 PCB$V_ASTPEN 00000011 PCB$V_BATCH 0000000E PCB$V_DELPEN 00000001 PCB$V_DISAWS 00000018 PCB$V_FORCPEN 00000002 PCB$V_HIBER 00000013 PCB$V_INQUAN 00000003 PCB$V_LOGIN 00000014 PCB$V_NETWRK 00000015 PCB$V_NOACNT 0000000F PCB$V_NODELET 00000017 PCB$V_PHDRES 00000012 PCB$V_PSWAPM 00000004 PCB$V_PWRAST 00000016 PCB$V_RES 00000000 PCB$V_RESPEN 00000005 PCB$V_SSFEXC 00000006 PCB$V_SSFEXCE 00000007 PCB$V_SSFEXCS 00000008 PCB$V_SSFEXCU 00000009 PCB$V_SSRWAIT 0000000A PCB$V_SUSPEN 0000000B PCB$V_SWPVBN 00000010 PCB$V_WAKEPEN 0000000C PCB$V_WALL 0000000D PCB$W_APTCNT 00000030 PCB$W_ASTCNT 00000038 PCB$W_BIOCNT 0000003A PCB$W_BIOLM 0000003C PCB$W_DIOCNT 0000003E PCB$W_DIOLM 00000040 PCB$W_GPGCNT 00000034 PCB$W_GRP 0000008A 31 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL PCB$W_MEM 00000088 PCB$W_MTXCNT 0000000E PCB$W_PPGCNT 00000036 PCB$W_PRCCNT 00000042 PCB$W_SIZE 00000008 PCB$W_STATE 0000002C PCB$W_TMBU 00000032 displaying which processes are running in the system One can obtain a list of the processes in the system using the SDA command SHOW SUMMARY The results of and SDA show summary command are show below PID Process name Image name State Pri UIC WKSET --- ------------ ---------- ----- --- --- ----- 00010000 NULL COM 0 [000,000] 0 00010001 SWAPPER HIB 16 [000,000] 0 00010050 USEFUL_WORK SYS$SYSROOT:[SYSEXE]SDA.EXE;1 CUR 5 [002,002] 233 00060051 MIS41 SYS$SYSROOT:[SYSEXE]COPY.EXE; LEF 5 [104,041] 67 00010052 AUTOLGOT SYS$SYSROOT:[SYSEXE]NEWPID.EXE HIB 9 [015,001] 39 00010053 DRD0BACP SYS$SYSROOT:[SYSEXE]F11BACP.EX HIB 13 [001,003] 179 00170054 _TTA1: LEF 5 [002,011] 60 00010055 OPCOM SYS$SYSROOT:[SYSEXE]OPCOM.EXE; LEF 8 [001,004] 79 00010056 JOB_CONTROL SYS$SYSROOT:[SYSEXE]JOBCTL.EXE HIB 9 [001,004] 311 00020057 SYSWARREN SYS$SYSROOT:[SYSEXE]BACKUP.EXE LEF 5 [002,011] 250 00010058 DRA0BACP [SYSEXE]F11BACP.EXE HIB 9 [001,003] 410 00020059 ERRFMT SYS$SYSROOT:[SYSEXE]ERRFMT.EXE HIB 7 [001,006] 59 PID The first column is the PID. For more information on the PID see section 2.1.5 . 32 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL PROCESS NAME The next column is the process name. This information is from the PCB. The Process name is not necessarily unique; however only one process per group may have any given name. IMAGE NAME The image name if available is the name of the program the process is running, if it is running a process; otherwise, it will be blank. This information is is contained in mmg$imghdrbuf58 for the current process. STATE The state is the next column. For information about process states see section 2.1.7 PRIORITY The next column is the current priority. For information about current priority see section 2.1.8 UIC The next column is UIC. This is the user indentification code and is usually mainly in conjuntion with RMS. 33 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL WKSET The next column gives the number of pages in the working set. This information is obtained from the process header. The offset is defined as PHD$L_WSSIZE. DISPLAYING THE ABOVE INFORMATION USING CCL To display the above information using CCL one starts by examining SCH$GLPCBVEC, and SCH$GLMAXPIX. One also needs the value of SCH$GLNULLPCB. The values of these for version three are SCH$GL_PCBVEC = 800026A4 SCH$GL_MAXPIX = 800026A8 SCH$GL_NULLPCB = 80002350 the figure below shows the lists of examines one does and the system respones (the numbers are for reference). 1 >>>H The computer responds: HALTED AT 80007B06 2 >>>E/V 800026A4 The computer responds: P 003A9EA4 800E040C 3 >>>E/V 800026A8 The computer responds: P 003A9EA8 00000059 4 >>>E/V 800E040C The computer responds: P 003CD610 80002350 5 >>>E The computer responds: P 003CD614 80002568 6 34 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL >>>E The computer responds: P 003CD618 80002350 7 >>>E The computer responds: P 003CD61C 80002350 . . . 8 >>>E The computer responds: P 003CD754 800EFB80 9 >>>E The computer responds: P 003CD758 800F1200 10 >>>E The computer responds: P 003CD75C 800EC3E0 11 >>>E The computer responds: P 003CD760 800F40E0 12 >>>E The computer responds: P 003CD764 800EA4A0 13 >>>E The computer responds: P 003CD768 800ECE80 14 >>>E The computer responds: P 003CD76C 800EB080 15 >>>E The computer responds: P 003CD770 800F4860 16 35 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL >>>E The computer responds: P 003CD774 800ED800 17 >>>E The computer responds: P 003CD778 800F4D60 NOTE In the rest of this example the physical addresses will be omitted. However in actually doing this the computer would give them. 18 >>>E/V 800023B0 The computer responds: 00010000 19 >>>E/V 800023B8 The computer responds: 4C554E04 20 >>>E The computer responds: 0000004C 21 >>>E/V/W 8000237C The computer responds: 000C 22 >>>E/V/B 8000235B The computer responds: 1F 23 >>>E/V 800023D8 The computer responds: 00000000 24 >>>E/V/W 80002386 The computer responds: 0000 . . . 36 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL 25 >>>E/V 800F4140 The computer responds: 00010053 26 >>>E/V 800F4148 The computer responds: 44524408 27 >>>E The computer responds: 43414230 28 >>>E The computer responds: 00000050 29 >>>E/V/W 800F410C The computer responds: 0007 30 >>>E/V/B 800F40EB The computer responds: 12 31 >>>E/V 800F4168 The computer responds: 00010003 32 >>>E/V/W 800F4114 The computer responds: 0040 33 >>>E/W The computer responds: 0073 37 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL explanation of the above 1 The first command halts the computer. this is necessary before doing CCL examines. 2 here we are examining the location SCH$GL_PCBVEC. This location points to the vector of process control blocks; i.e. starting in location 800E040C are pointers to each of the PCBs. The index part of the PCB is an offset into this vector. 3 Next we examine SCH$GL_MAXPIX. The contents of this location is the length of the vector mentioned above. It is also the value of the sysgen parameter MAXPROCESSCNT. This tells us how many examines we must do to find all the PCBs. 4 The next examine looks at the first entry in the PCB vector. it is a pointer to the null process. In the PCB vector this value will be in all unused locations; it will however always be in location 0. In the later examines we will compare what we come up with to this value to determine whether we have another process or not. 5 The next examine yield us a new PCB namely the PCB for the SWAPPER. 6 This examine finds the value 80002350 again so we know this location is unused. 7 again we find 80002350. The elepsis indicates that all these examines also yield this value; since VMS normally uses the first two slots and then fills slots in from the other end. 38 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL 8 This is the address of the PCB for the process USEFUL_WORK. 9 This is the address of the PCB for the process MIS41. 10 This is the address of the PCB for the process AUTOLGOT. 11 This is the address of the PCB for the process DRD0BACP. 12 This is the address of the PCB for the process _TTA1:. 13 This is the address of the PCB for the process OPCOM. 14 This is the address of the PCB for the process JOB_CONTROL. 15 This is the address of the PCB for the process SYSWARREN. 16 This is the address of the PCB for the process DRA0BACP. 17 This is the address of the PCB for the process ERRFMT. At this point we have the addresses of all the PCBs of processes in the machine. We are now ready to display information about these processes. Since this is a long task, only two processes are shown; however, the technique would work for all of them. Normally when displaying information about the system using CCL, one would have some idea of what information one needed and only bother to display that. 39 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL 18 This is the PID of the NULL process. The number 800023B0 is obtained by taking 80002350 and adding PCB$L_PID which is 60. 19 This is the beginning of the PROCESS NAME. The number 800023B8 is obtained by taking 80002350 and adding PCB$T_LNAME which is 68. The first byte of this field is a 4 which tells us that the name is four characters long. The next three bytes are the ASCII for NUL. 20 This gives us the last L. 21 This gives us the STATE. The number 8000237C is obtained by taking 80002350 and adding PCB$W_STATE which is 2C. The value C indicates COM (see 2.1.7). 22 This is the current priority. The number 8000235B is obtained by taking 80002350 and adding PCB$B_PRI which is B. Note that internally priorities are coded with 1F being the lowest and 0 being the highest. To obtained the priority that would be displayed this number must be subtracted from 1F. 23 This is the UIC. The number 800023D8 is obtained by taking 80002350 and adding PCB$L_UIC which is 88. 24 This is the number of physical pages that the process occupies. The number 80002386 is obtained by taking 80002350 and adding PCB$W_PPGCNT which is 36. 25 This is the PID of the DRD0BACP process. The number 800F4140 is obtained by taking 800F40E0 and adding PCB$L_PID which is 60. 26 40 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL This is the beginning of the PROCESS NAME. The number 800F4148 is obtained by taking 800F40E0 and adding PCB$T_LNAME which is 68. The first byte of this field is an 8 which tells us that the name is eight characters long. The next three bytes are the ASCII for DRD. 27 This is 0BAC. 28 This is the Final P. 29 This gives us the STATE. The number 800F410C is obtained by taking 800F40E0 and adding PCB$W_STATE which is 2C. The value 7 indicates HIB (see 2.1.7). 30 This is the current priority. The number 800F40EB is obtained by taking 800F40E0 and adding PCB$B_PRI which is B. Note that internally priorities are coded with 1F being the lowest and 0 being the highest. To obtained the priority that would be displayed this number must be subtracted from 1F. 31 This is the UIC. The number 800F4168 is obtained by taking 800F40E0 and adding PCB$L_UIC which is 88. 32 This is the number of global pages that the process is using. The number 800F4114 is obtained by taking 800F40E0 and adding PCB$W_GPGCNT which is 34. 33 This is the number of physical pages that the process occupies. The number 800F4116 is obtained by taking 800F40E0 and adding PCB$W_PPGCNT which is 36. 41 PAGESWAPPER - December 1983 - Volume 5 Number 6 Using SDA and CCL MODIFYING THE RUNNING SYSTEM Modifying the priority of the current process In order to modify the priority of the current process (from CCL) one must: make sure one is in kernel mode find the location of the process PCB modify the process' current priority modify the process' base priority modify the system stored current priority To do this one first halts the cpu. Then one examines the PSL by typing E PSL Next one gets into kernel mode by typing D PSL 0 Next one locates the process' PCB by doing E/V 8000210C or whatever the current value of SCH$GLCURPCB is One then takes this value n adds B to it to get the location of the current priority and does a Deposit. If the value n was 800F4FE0 and one want to change the current priority to 1 (which is represented 1E) one would D/V/BYTE 800F4FEB 1E Next one modifies the base priority NOTE ONE MUST BE SURE TO MAKE THE BASE PRIORITY LOWER THAN OR EQUAL TO THE CURRENT PRIORITY by adding 2F to n and deposity a number in that location. In the above example if we wished to set the base priority to 0 we would D/V/BYTE 800F500F 1F The last thing that must be done is to modify SCH$GB_PRI. given its current value for version three to set it equal to the current priority one would D/V/BYTE 8002158 1E 42 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT INPUT/OUTPUT A SIG Information Interchange A form for INPUT/OUTPUT submissions is available at the back of the issue. INPUT/OUTPUT 207 Caption: LISP & LISP Compilers & LISP-simili Message: We are interested in Lisp compilers or interpreters running on VAX 11/780 VMS 3.3 and in exchanging LISP programs and experience (for example - word processing, language analysis for translation, VT125 or Tektronix Plot 10 graphics). Contact: Maria Luisa Luvisetto INFN-CNAF Via Mazzini N.2 40138 Bologna Italia 051/498111 - TELEX 510590 INPUT/OUTPUT 208 Caption: Micom - Borer 600 switch Message: We would like to use the above switch to transfer some files from our PDP-11 to our VAX. If anyone has had experience transferring data using this device between any combination of VAXs or PDPs, please contact me. Contact: Robert Palmer Needham & James Windsor House Temple Row Birmingham, England 021 236 9701 Date: September 9, 1983 43 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT INPUT/OUTPUT 209 Caption: G-Floating hardware information Message: We have a large program that we are converting from IBM to VAX, and because of the size of some of the numbers, we must use G-floating representation. Our tests have shown that this takes 70 to 100 times as much CPU time as D-floating with the floating point accelerator, because it is emulated in software. We are considering buying the G-floating hardware and we have so far been unable to find out how much faster it is than the emulator. Does anyone have any experience or knowledge of this? Contact: Elizabeth Buie Computer Sciences Corporation 8728 Colesville Road Silver Spring, MD 20910 (301) 589-1545 ext. 596 Date: November 3, 1983 INPUT/OUTPUT 210 Caption: XAA0: Times Out Message: We have a "device time out" problem occurring on an intermittent basis. We talk between our 780 through DR11W to an AYDIN 5215 display generator. We (naturally) think our software is clean. Any help as to where to look next? Contact: G. E. Taylor Westinghouse Electric Corporation 777 Penn Center Boulevard Pittsburgh, PA 15235 (412) 824-9100, Ext. 209 Date: November 4, 1983 44 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT INPUT/OUTPUT 211 Caption: Need display utility Message: We need a program that will allow for viewing on a VT100 terminal a corporate report, length/page of 66, width of 132 with up, down, top, bottom and find functions. The EDT editor is not acceptable because the definition of the function keys cannot be deisplayed on the same screen as the report. We have a VAX 11-750 with Basic, Cobol, Fortran. Contact: Joyce Nelson 845 North 35th Street, Room 106 Milwaukee, Wis 53208 (414) 678-3205 Date: November 8, 1983 INPUT/OUTPUT 212 Caption: VAX/VMS 7 track tape drive support Message: We are looking for a device driver and diagnostics for a TU10 in 7 track mode. Contact: John F. Stitzinger, Engineer HRB Singer D/115 Science Park RD PO Box 60 State College, PA 16801 (814) 238-4311 x2853 Date: November 10, 1983 INPUT/OUTPUT 213 Caption: File Archiving REPLY TO I/O # 186 Message: Found software which (according to vendor) 1. maintains disk directory of user tapes/files, 2. issues mount requests to operator via SYS$TAPE batch queue, 3. performs reads/writes when mounted, 4. requests/releases tapes, 5. allows ANSI, DOS, NOLABEL, FOREIGN ... attributes, 6. maintains location info for operator and allows add, remove, reassign, resues of tapes by operator. We were going to buy it but our budget got crunched. Get info from International Structural Engineers, P.O. Box 24073, Los Angeles CA 90024-0073 (213) 837-8339. 45 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT From: Bill Adkins Institute for Mining and Mining Research PO Box 13015 Iron Works Pike Lexington, KY 40512 (606) 252-5535 x224 Date: November 15, 1983 INPUT/OUTPUT 214 Caption: Magnetic Tape Input Software Message: We need a program to read ASCII or EBCDIC files from an unlabelled tape with multiple files of different blocking. We have a VAX 11/780 running VMS with a TE16 and a TU45. Contact: Gary Zouzoulas 801 North First Street, Room 500 San Jose, CA 95110 (408) 277-4031 Date: November 15, 1983 INPUT/OUTPUT 215 Caption: Editing below the '[EOB]' in EDT - Help! Message: Occasionally, during an EDT editing session I have found my cursor to have jumped below the '[EOB]'. At this point I may have one to five scrolling lines below the [EOB] and one scrolling line at the top of the terminal screen. This has occured on VT-100's with and without advanced videa, a VT-125, and on two separate VAX 11/780's. What causes it, how do you prevent it, and how do you stop it (aside from ). Contact: Diane Harris Computer Sciences Corporation 443 Inyokern Road Ridgecrest, CA 93555 (619) 375-4898 Date: November 15, 1983 46 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT INPUT/OUTPUT 216 Caption: Need Retail Bank Application System Message: Interested in Purchase/License and cooperating in enhancement. Contact: Ray Evanson Paragon Data Systems 927 East King Street Winona, MN 55987 (507) 452-8833 Date: November 16, 1983 INPUT/OUTPUT 217 Caption: RL01's on a VAX Message: We have 3 RL02's and a single RL01 commected via an RL211 controller to a VAX-11/780. VMS recognizes and is quite happy with the RL01 (and does not think it is an RL02 but truly a 5 Mbyte RL01). However, field service does not support the RL01 since the RL02 diagnostics do not recognize it as such (they try to think it is an RL02). Contact: Frank J. Nagy Fermilab P. O. Box 500 MS/306 Batavia, IL 60510 (312) 840-4935 Date: November 16, 1983 INPUT/OUTPUT 218 Caption: Setting Top of Form on LXY Printers Message: I have found a very fast way to set TOF after a paper jam or running out of paper on an LXY. Following these simple steps will save you time, guaranteed: When your paper is properly adjusted at TOF, mark the location where the first perforation above the tractor feeds lays against the printer cabinet (approximately even with the hinges under the cabinet door). Now, whenever you have to realign paper to TOF: 47 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT 1. Without paper in the tractor feeds, depress the TOF button (red check light must be off). 2. Align paper into tractor feeds so that a perforation is lined up with the mark you made by following instructions above. 3. Set the platen thickness knob appropriately, close the cabinet door and set the printer on-line. Contact: richard N. Eliopoulos 5258 Vickie Drive San Diego, CA 92109 (619) 225-7426 Date: November 16, 1983 INPUT/OUTPUT 219 Caption: Producing breaks Message: We need to be able to generate breaks with a DZ-11 or DMF-32. Even though the hardware can do it, VMS doesn't appear to support the sending of breaks. We're a bit reluctant to patch our device drivers, and know about the kludge of sending 50 baud NULLs to simulate a 9600 baud break. Contact: Sam Sjogren Caltech 356-48 Pasadena, CA 91125 (213) 356-6660 Date: November 17, 1983 INPUT/OUTPUT 220 Caption: Looking for LPA Sites Message: I am looking for sites doing real-time data input with the LPA sybsystem on a VAX 730 or 750. I would like to talk to sites using this subsystem. Contact: Fred D. Brandt SCRL-MTB Gallandet College 7th + Fla Avenue NE Washington, DC 20002 (202 651-5439 48 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT Date: November 18, 1983 INPUT/OUTPUT 221 Caption: Type command Message: We need to have information on the availability of software which would allow the qualifier /LA=n to be used with the type command in order to limit the viewing of a large file to the last few pages or lines as needed. This would be similar to the TOPS-10 type command. Contact: Dave DiGennaro Coopers & Lybrand One Post Office Square 28th Floor Boston, MA 02109 (617) 574-5280 Date: October 19, 1983 INPUT/OUTPUT 222 Caption: Lots of terminals - REPLY TO I/O # 198 Message: While we haven't connected > 64 terminals to a VAX, we do hook up 40 ports. Our experience has been (in a mostly software development environment) that we get a maximum of about 22 concurrent users on a 780 before degradation sets in. With more users than that, response time degrades rapidly. My point is that while it is physically possible to connect 64 terminals to a VAX 11/780, it may not be practical from a load viewpoint. Contact: Bob Ventimiglia Sanders Associates, Incorporated 95 Canal Street Nashua, NH 03063 (603) 885-5454 Date: November 5, 1983 49 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT Submission Form INPUT/OUTPUT Submission Form A SIG Information Interchange Please reprint in the next issue of the Pageswapper If this is a reply to a previous I/O, which number? ________ Caption: ______________________________________________________ Message: ______________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ ________________________________________________________________ Contact: Name _______________________________________________________ Address ____________________________________________________ ____________________________________________________________ ____________________________________________________________ Telephone ____________________________ Signature _____________________________ Date ________________ Mail this form to: PAGESWAPPER Editor, DECUS, MRO2-1/C11, One Iron Way, Marlborough, MA 01752, USA 50 PAGESWAPPER - December 1983 - Volume 5 Number 6 INPUT/OUTPUT Submission Form Tear out to submit an I/O item PAGESWAPPER Editor DECUS, MRO2-1/C11 One Iron Way Marlborough, MA 01752 USA 51 PAGESWAPPER - December 1983 - Volume 5 Number 6 System Improvement Request Submission Form System Improvement Request Submission Form Page 1 of _____ ________________________________________________________________ Submittor: Firm: Address: Phone: ________________________________________________________________ How to write an SIR: Describe the capability you would like to see available on VAX systems. Be as specific as possible. Please don't assume we know how it's done on the XYZ system. Justify why the capability would be useful and give an example of its use. If you wish, suggest a possible implementation of your request. ________________________________________________________________ Abstract (Please limit to four lines): ________________________________________________________________ Description and examples (use additional pages if required) 52 PAGESWAPPER - December 1983 - Volume 5 Number 6 System Improvement Request Submission Form Tear out to submit an SIR Gary L. Grebus Battelle Columbus Laboratories 505 King Avenue Columbus, Ohio 43201 USA 53