1993 HP1000 Guru Column Q. I am using a 35401 Autochanger on my RTE-A system to backup onto multiple tapes. Is there a way with FST to have it distribute the LU's I'm backing up onto specific tapes? For instance, I would like to put LU's 16 and 17 onto tape 1, then 18 and 19 onto tape 2....and so on. A. You can use the FST MT command as follows to accomplish such a task: FST> mt, lu * Specify the 35401 LU (Load first tape) FST> y * Overwrite without asking FST> ba 16 * Backup LUs 16 and 17 FST> ba 17 FST> go FST> mt * Unload the first tape and FST> mt, lu * load the next tape FST> un * Unselect previous backups FST> ba 18 * Backup LUs 18 and 19 on the next tape FST> ba 19 FST> go FST> ex This method will allow you break up your backup tapes any way you like. Q. Do the 7957A and the 7957B disk drives have the same capacity, IE number of blocks? I never seem to be able to find a complete list of the block sizes of all the disks I use. A. This is a very common question, so we'll try to provide a complete list of all CS80 and new SCSI disks here. Disk Useable RTE blocks Approximate size ----------------------------------------------------------------- 7907A 80,256(rem) + 80,256 (fixed) 41 Mbytes 7908A 64,750 16 Mbytes 7911P/R 109,824 28 Mbytes 7912P/R 256,256 65 Mbytes 7914P/R 516,096 132 Mbytes 7933H/XP 1,579,916 404 Mbytes 7935H/XP 1,579,916 404 Mbytes 7936H/XP 1,201,956 307 Mbytes 7937H/XP 2,232,204 571 Mbytes 7941A 92,928 24 Mbytes 7942A 92,928 24 Mbytes 7945A 216,832 55 Mbytes 7946A 216,832 55 Mbytes 7957A 319,095 81 Mbytes 7957B 319,788 82 Mbytes 7958A 510,552 130 Mbytes 7958B 594,216 152 Mbytes 7959B 1,188,432 304 Mbytes 7961B 319,788 82 Mbytes 7962B 594,216 152 Mbytes 7963B 1,188,432 304 Mbytes 9122C 4,928/floppy (2) 1.2 Mbytes 9122D 2,464/floppy (2) 630 Kbytes 9133D 58,176 / 2,464 15 Mbytes / 630 Kbyte 9133H 78,336 / 2,464 20 Mbytes / 630 Kbyte 9133L 156,320 / 2,464 40 Mbytes / 630 Kbyte 9134D 58,176 15 Mbytes 9134H 78,336 20 Mbytes 9134L 156,320 40 Mbytes 9153B 77,808 / 2,464 20 Mbytes / 630 Kbyte 9153C option 10 39,088 / 4,928 10 Mbytes / 1.2 Mbyte 9153C option 11 39,088 10 Mbytes 9153C option 20 78,624 / 4,928 20 Mbytes / 1.2 Mbyte 9153C option 21 78,624 20 Mbytes 9153C option 40 157,248 / 4,928 40 Mbytes / 1.2 Mbyte 9153C option 41 157,248 40 Mbytes 9262B 594,216 152 Mbytes 9263B 1,188,432 304 Mbytes C2200A (335H) 1,309,896 335 Mbytes C2202A (670H) 2,619,792 670 Mbytes C2203A (670XP) 2,619,792 670 Mbytes SCSI disks RTE blocks (*) Approximate size ------------------------------------------------------------------------ C1701A (650/A MO) 1,147,248 (92279A media) 294 Mbytes/side " " " " 1,258,272 (92280A media) 322 Mbytes/side C2212A 1,296,512 330 Mbytes C2213A 2,593,024 660 Mbytes C2214B 5,294,160 1355 Mbytes C2216T 2,645,280 677 Mbytes C2217T 5,294,160 1355 Mbytes C2235 1,649,592 422 Mbytes C2295B 5,294,160 1355 Mbytes C2460F/R 1,649,592 422 Mbytes C2461F/R 2,645,280 677 Mbytes C2462F/R 5,294,160 1355 Mbytes C2472S 1,649,592 422 Mbytes C2473S 2,645,280 677 Mbytes C2474S 5,294,160 1355 Mbytes C2481A 2,645,280 677 Mbytes C2482A 5,294,160 1355 Mbytes * Note: These are the "RTE" blocks available. All of the above SCSI discs have a blocking factor of 2, which means the physical disc sector size is 512 bytes, which is equal to two 256 byte "RTE" blocks. Specify DP:8:1 in your DVT for these discs. Q. I have heard that SCSI disks can be genned with a larger size then the current limit of 32,767 RTE tracks. Is this true? A. This is true. The SCSI disk driver allows up to 65,535 tracks on an LU while the HPIB disk driver is limited to 32,767 tracks per LU. However, the extra tracks cannot be used with the RTE-A file system. A disk LU with more than 32,767 tracks cannot be mounted as a CI volume nor as a FMGR cartridge. The larger disk LUs can only be used by applications which use exec calls to access the disk driver directly and bypass the file system. Listed below are the size limits imposed by the SCSI and HPIB disk drivers and by the RTE-A file system. Tracks Blocks per LU per Track Blocks Bytes ------ --------- --------- ----------- File System 32,767 64 2,097,088 536,854,528 DD*33 32,767 128 4,194,176 1,073,709,056 DDQ30 65,535 128 8,388,480 2,147,450,880 So, while the disk drivers are capable of handling larger LU sizes, the file system currently limits the disk LU size to just under 537 Mbyte Q. Is it possible to put the 24612A A Series diagnostics into the reserved area of a CI volume? A. Yes, as of rev 5270, this is possible. FPUT is used to install bootable systems (BOOTEX) into the reserved area of a CI volume. At 5270, FPUT was enhanced to allow installing the 24612A CPU diagnostics into the reserved area. This was accomplished by adding a second offset, offset1, to the FPUT runstring. FPUT's runstring is now as follows: FPUT,filename,lu,[offset0],[offset1] where: filename = Bootable file lu = disc volume [offset0] = VCP file (256 block) offset from start of disk volume [offset1] = 32 block offset into [offset0] Because the 24612A CPU diagnostics are based on a 32 block file structure, FPUT needed to be able to concatenate these files on 32 block boundaries within the 256 block VCP file. The 24612A diagnostics are controlled by what is known as BCM, or Basic Control Module. For example, to install the 24612A diagnostics your commands would look something like this: FPUT,BCMSC,16,3,0 puts BCM after a bootex, into VCP file 3 FPUT,#AUTO,16,3,2 puts #AUTO after 64 block BCM FPUT,%MAPS,16,3,3 puts %MAPS after 32 block #AUTO FPUT,%MSGS,16,3,4 puts %MSGS after 32 block %MAPS FPUT,%DDL, 16,3,5 puts %DDL after 32 block %MSGS Note that the files are either 64 or 32 blocks in size, and offset(1) is set accordingly. The 24612A diagnostic files are typically shipped with the Primary system, and starting at revision 5270, a command file was supplied for installing the diagnostics into a reserved area of the bootable CI volume. At 6.0, the Primary installation will allow you to have the diagnostics installed automatically as the Primary is restored. And, oh, by the way, at 6.0, the RTE-A Primary will no longer be in ASAVE format, it will be in FST format. More on this later, or check out the 6.0 Primary Installation manual! Q. I have a 1200/2400 baud modem on a D mux. How can I have the mux change baud rates for each login? The way I have it now, once someone logs on at 1200, that's what it stays set to. I want it to revert back to speed sense each time. A. The way to do this is with a undocumented bit in the 30B control request. Setting bit 7 will enable speed sense at logon. For example, to have port 7 do speed sense at each logon, you would initialize the port with the following 30B command: CI> CN,lu,30B,377B Now the mux port will re-speed sense at each logon. SPORT will look as follows: Status for LU 127: Device Driver: DDC01 Rev. 6000 Driver type = 05 Interface Driver: ID800 Rev. 6000 Firmware: Rev. 5.20 CN20: Primary Program: PROMT (Enabled) CN40: Secondary Program: HPMDM (Disabled) CN17: 000000B No user defined terminators CN22: 30000 Timeout = 300.00 seconds CN30: 000377B Frame=8/1 None BRG0 Speed Sensing Port 7 SS at Logon CN31: 000000B CN33: 000000B CN34: 000002B HP Protocol DV20: 000077B Character mode DVT Address: 52604B; IFT Address: 57632B Of course, your available baud rates are still restricted to the triplets defined by a 16B command or the default triplet after the first 30B, or speed sense on the port. Q. Speaking of triplets on the mux, it sure would be nice to KNOW what triplet the baud rate generator is using. How do I know? A. Patience will work in your favor here. The D mux driver is being enhanced to allow saving the triplet being used at initialization and this information will be available in the SPORT listing. Stay tuned. Q. What gives with these SCSI default gen records? If I gen a SCSI disk with M64Mb as the default, the resulting disk LU is really over 67 Mbyte. Why call it a 64Mb disk? A. A subtle change occured here, and you will see this change in the 6.0 version of FREES. It has to do with the definition of "Mb". If you are an old hardware type, then you are used to disk drives having 20 Mbyte or 120 Mbyte, which meant Millions of bytes. But if you had a 64 Kbyte RAM memory board, that was REALLY 65,536 bytes. Well, apparently we have decided that we will refer to disk sizes the same way. And FREES at 6.0 will report the size in Mbytes, as opposed to the true "millions of bytes". So when you use M64Mb as the gen record, you are really allocating 64 x 1024 x1024 = 67,108,864 bytes. This is critical as you get towards the end of the disk. You can't simply add up the gen record sizes, and compare them to the disk size specified in the disk documentation, it won't fit. For example, if you are genning one of the new 1.05 Gbyte discs, you cannot gen in eight 128Mb LU's. Although 8 x 128 = 1,024 Mbyte, it won't fit because it is REALLY 1,024 x (1024 x 1024) = 1,073 million bytes. You are about 20 Mbyte over. There may be an enhancement to FREES later to give a choice of reporting either MBytes, or "Millions of bytes". Q. I am troubleshooting a MAIL/1000 problem, so I OF'ed RMAIL. But when I ran MAIL and created a mail message and sent it, RMAIL kicked in and sent it. How do prevent RMAIL from sending messages I don't want sent? A. Easy. Just Suspend RMAIL. Now you can create all the messages you want and MAIL them, and they will remain in the /MAIL/QUEUE directory until you either PURGE the files or GO RMAIL to send them. This method allows you to create messages and examine or delete them at will. Q. Is there a way to set or clear the backup bit on a file without actually backing it up? A. At 6.0 there is. Use TOUCH. TOUCH is almost the same as the UN*X touch, and allows setting or clearing the backup bit on any file or files selected via standard masking. TOUCH also allows updating the timestamps of files. Checkout the TOUCH help file for more details. Q. I have a user who is complaining about not receiving MAIL messages, yet when I check the TRAFFIC.LOG, I see messages for the user but they don't say "delivered". What's going on here? A. I think your user is pulling a fast one on you. Here's a TRAFFIC.LOG that shows the failure: Sep 29 3:54 pm <9209291553.ZM9021@hpwrch.mayfield.hp.com>: 514 bytes From geff@hpwrch.mayfield.hp.com to walt@hpwrcxe.mayfield.hp.com Sep 29 3:56 pm <9209292255.AA09038@hpwrch.mayfield.hp.com>: 698 bytes From geff@hpwrch.mayfield.hp.com to walt@hpwrcxe.mayfield.hp.com Note that neither of these messages actually shows delivered, nor routed elsewhere. These messages simply vanished. And the reason is an entry in the FILTER.RC file for user "walt" that says: from * : refuse This acts as a "filter" and refuses messages from "*" , which is a wildcard (like the @ for filemasks) that means "All Users" So in this case, no MAIL will be delivered to "walt". The usefulness of this command would be to turn off MAIL from your manager, if you are tired of being pestered about that overdue project. You can always claim innocence and say the mail must have gotten lost! Q. I'm confused by the difference between the owner of a global directory as reported by OWNER versus FOWN . OWNER reports a different owner than FOWN. What's the difference? A. At revision 5.0, volume ownership was introduced, in addition to the normal directory ownership. This caused a change in the owner reported by OWNER versus FOWN. OWNER will always display the true owner of directories. FOWN will display the following: - Global directories will show as being owned by the OWNER of the volume. - Subdirectories will show as being owned by the OWNER of the global directory The primary function of FOWN is to display disc usage for owners. To display the true owner, always OWNER. Q. I'm having trouble setting up a printer on my D MUX to handle both carriage control (via column 1) and CR-LF suppression with the underscore character at the end of the records. How is this accomplished. A. This is done with a somewhat poorly documented CN 34B parameter. Bit 12 of the CN34B, the "nice" bit, also will configure DDC12 compatibility mode. This allows both carriage control and CR-LF suppression. So for a HP printer using XON-XOFF, you could use CN, , 34B , 14401B This will give you Pre-spacing Printer Xon/Xoff Protocol, with the "nice" bit set for DDC12 compatibility. Now a _ (underscore) at the end of a record will suppress CR-LF. Q: I'm confused about the A900 firmware and the correct RPL file. I thought I had it clear, but the 6.0 Communicator on page 4-7 mentions something about "pre 4.0 RPL files". What are these? A: Actually, this is a typo. What it should say is pre 4.0 FIRMWARE. Prior to revision 4.0 of the A900 firmware, the cross move byte instructions were broken, so they were commented out of the A900 RPL files. This necessitated the use of XMB.REL to avoid undefined references to MB10, MB12 and MB21. Now that the firmware has been fixed, instead of using XMB.REL, you can use RPL_A900_REV4.REL. You must still use the same RPL90 or RPL91 as before, RPL_A900_REV4.REL only contains the cross move byte instructions. Q: How can I tell if I have the revision 4 A900 firmware? A: The easiest way is to look at the ROMs themselves on the A900 Sequencer card and compare them to the list on page 3-122 in the 6.0 Communicator. Table 1 also lists the 4.0 firmware. Table 1 12201-80103 (U0803)* 12201-80104 (U0802)* 12201-80105 (U0801)* 12201-80106 (U1103)* 12201-80107 (U1102)* 12201-80108 (U1101)* 12201-80109 (U0703)* 12201-80110 (U0702)* 12201-80111 (U0701)* 12201-80112 (U1003)* * This firmware is available from HP 12201-80113 (U1002)* as part number 12201-60103 12201-80114 (U1001)* 12201-80115 (U0603)* 12201-80116 (U0602)* 12201-80117 (U0601)* 12201-80118 (U0903)* 12201-80119 (U0902)* 12201-80120 (U0901)* 12201-80042 (U1407) 12201-80043 (U1607) Q: I am using the 6.0 TOUCH program and it is Memory Protecting or it just hangs. What gives? A: Turns out the TOUCH program has a bug that causes it to fail on all A400, A600 and A700 processors. TOUCH works correctly on the A900 and the A990. A patch is available from the Response Center, or contact your local HP representative. Q: I have some literature that refers to a company called COMCSI Data Systems. They have a product called IMF/1000. I have been unable to contact this company. Are they still around? A: Yes. They are located in Atlanta, Geogia. PO Box 500595, 31150 Their phone number is (404) 552-9499. IMF/1000 is a Image Maintenance Facility. Please be aware that HP does not support IMF, and cannot endorse it's use. Q: I am using HPMDM to set up my modems, but when I try to change the default 1200 baud and HP protocol, HPMDM seems unable to "do it". What gives? A: Unfortunately, this is a problem with HPMDM at 6.0. All the option commands, like HA=XO, BA=9600 do not work. This has been fixed, and a patch is available from the Response Center or your local HP representative. Q: I just read in my Programming manual that I can now take advantage of of the new Extended VMA feature of the A990 if I have revision 10 (decimal) of the A990 firmware. My HP Customer Engineer just installed the current EPROM, part number 12990-80111, which is reported as revision 9 (decimal). Is there a newer ROM? A: No. Revision 10 is not yet available. However, with revision 9, and RTE-A revision 6.0, you can use the DOWNLOAD program to modify the instruction set microcode in the A990 Writeable Control Store and add the new Extended E/VMA features. The DOWNLOAD program also copies the microcode file to a SHEMA partition in main memory. This file is used to restore the microcode in the event of a power failure. You can use the A990FWID program to determine the revision of the A990 EPROM. The following is the first few lines of the programs output, the revision of the whole EPROM is the first product: (Note that A990FWID will not load on a 5270 system) Product Number Supercode/Revision Product Description (in octal) (in octal) -------------- ------------------ --------------------------- 000 000011 Entire EPROM 001 000024 VCP 002 000002 STST0 (Self-test module 0) Q: I have a Magneto-Optical disk pack that someone gave me. I need to recover the data, but the track map of the original system does not match my system. Is there a easier way to modify my track map other than regenning? A: Yes. Both the SCSI and the CS80 drivers, DDQ30 and DD*33 have a 76B control request that allows modifying the driver parameters for a disk LU. The format of the request is: CN,lu,76B,p1,p2,p3,p4 Where: p1 = number of tracks (DP6) p2 = number of 128 word blocks per track (DP7) p3,p4 = starting block number(double integer) (DP4,DP5) This change only affects the tables in MEMORY. Once you reboot the system, the values will return to the genned values. Beware when changing the track map online so that you do not inadvertently reinitialize a "good" data cartridge. Q: Is it possible to copy a FST disk archive file to tape, and then use FST to recover individual files from the tape without restoring the entire archive file back to disk? A: Yes, this can be done. Since the first SCSI release of FST (5021), FST has had the capability to save files to a type 1 disk archive file. This file can be copied to tape, using CI, and then FST can be used to recover individual files from the tape. But because the disk archive file is a type 1 file with 128 word records, and FST reads in 5120 word chunks, we first must copy the archive file into a 5120 record length type 2 file. For example: CO,Fst_Archive_File,Type2:::2::5120 Then: CO,Type2,8 to copy to tape. Now FST can be used on the tape to recover individual files. As of FSTP revision 6002, FSTP now creates the archive file as a type 2 so the above step will no longer be necessary. If you want a copy of this patched FSTP, contact the Response Center. Q: I am using a 650/A Optical disk and noticed that even when being accessed, the drive can be ejected. I have set the S bit in the driver parameter 8. How do I prevent the disk from being being accidentally ejected during use? A: In order to implement spinup/spindown control, 3 things need to be done/verified: 1) You are using the CDS version of D.RTR 2) You have linked D.RTR with DDMAX.REL after modifying DDMAX.MAC to enable CN40/41 control 3) You have set BOTH the S and the E bits in DP8. This is due to bug in the driver. Thus DP8 for a 1024 byte sectored MO disk should be 140003b. Once you have done all the above, D.RTR will spin the drive up on a mount and down on a dismount. When mounted, it will also prevent media ejection via the front panel of the MO disk. Q: In the 6.0 documentation for GREP, I see reference to FGREP. I don't seem to have FGREP in my 6.0 programs. Am I missing something? A: GREP and FGREP are actually the same program file. During 6.0 installation, the command file /vcplus/crlinks.cmd will symbolically link (new feature of 6.0) various links between programs. These linked programs will execute the same program, but the programs behaviour and options may be different, based on the actual name the program is ID'ed as. Q: I have a program that works fine on an A900 system, but on the A990 at 6.0 I get the following error: CVIE aborted at address 17131 Unimplemented Instruction Current segment = 2 A = 100000 B = 53654 X=0 Y=0 E=0 O=1 WMAP=106002 INST=105247 Z=61364 Q=47251 CSMODE ON This worked fine at 5.27. The instruction is LBPC. What am I missing? A: The instruction LBPC did not exist in the 5.27 A990 RPL file, since it had not yet been implemented in firmware. At 6.0, LBPC is included in the A990 RPL file. In order to use the 6.0 LBPC instruction in firmware, you must either use the DOWNLOAD program with the REV10UPGRADE.MIC file, which will update the microcode dynamically (see page 4-4 of the 6.0 HP1000 Communicator) or install new A990 firmware. The firmware part number is 12990-80113. This will update the firmware to rev 12. Q: How can I set up the lp spooler at 6.1 to print directly from MAIL/1000? A: This is very easy. Once you have installed the lp spooler, in order to use it from within MAIL/1000, simply modify your $home/mail.rc configuration file. For example, here is a typical mail.rc file: ignore [default] set signature = /walt/mail/signature set visual on set justedit on set print lp -onb -olm3 -od %f set edit edit %f %# %n|4|s set prompt %m_%n> set record wastebasket set prefix > The line "set print lp -onb -olm3 -od %f" makes the lp spooler the default print utility for MAIL. Note that you may also need to add a -d if you have no default LPDEST set with lpadmin, or have not set and exported a value for LPDEST in your local environment. The "nb", "lm3" and "d" options are for No Banner, Left Margin indented 3 and Double sided printing. (Note that some of these options may not be valid for your printer. These are for a LaserJet 4si. Check the 6.1 Print and Spooling manual for info on LP. The LP spooler will work on a 6.0 system, but is not "officially" supported. Q: I am trying to restore an application database file from a FST backup tape. It starts restoring the file and after several minutes I get the following error: FMP ERROR : Illegal file position Cannot restore: UAR005:::2:11259:16 According to the FST directory, the file is 11,259 blocks and my disk has over 20,000 contiguous free blocks. What could be causing this? A: The directory entry on the FST tape could be misleading. You may in fact be runnning out of room on the target disk. In the above example the file has a MAIN size of 11,259 blocks. The FST directory does NOT show whether the file has any extents. If the file had 3 extents, for example, then the target disk would need over 44,000 free blocks AND this freespace would have to be allocated in 4 pieces of at least 11,259 blocks EACH in order to restore the file (because it is a type 2 file and type 2 files are restored exactly as they were, with extents). There is no way to tell on the tape how large the file actually is without restoring it. SR # 4701216051 has been filed to enhance FST to report a more meaningful error when this situation occurs. Q: Is there a way to have FST preconfigured for the tape LU, and other options so I don't have to enter these every time I run FST. I know I can have a transfer file, but is there something easier? A: Yes. At 6.0, FST can use a fst.rc file in your HOME directory, or in /CMDFILES that contains FST commands that are executed prior to any interactive or runstring processing. See the 6.0 Utilities manual page 2-5. Make sure you set PATH 0 to your HOME directory A note about a previous article concerning 76B control requests for CS80 and SCSI disks. I neglected to mention that these were new features of the driver at 5.27 or 6.0. In the future, I will attempt point out things that are NOT available at NOT current revisions!