SUNDSTRO@FINABO.BITNET.UUCP (02/09/87)
Help! Has anybody had the same problem that i have. When a users wants to backup his files on a TK50 tape. He uses the following command.With it he has twice got the system to crash in a middle of a saveset. The tape is left without the last two EOF marks but is has started on the saveset. EOF marks are no problem you have to put them there manually, but that takes approx. 1 hour (because the drive is so slow and you want to check that you are located on the correct location). I dont want to continously spend time puting EOF marks on tape. So can anybody tell if there are some errors in the command line or is it not correct to do it this way ( I can't see any errors (all is written on one line, I have edited it to two lines here)). As a note: the last time he tried saving the same files the second try was successfull. All savesetts have unique names. The command is as follows: $backup/ver dua0:<masu.saxen.run>betas1.*,check1.*,stat*.* - mua0:dtauql7mwts/save Or is this a so called "known bug" to every body else but me, if so is it fixed in 4.5. If its none of above have anybody any clues where it goes wrong and has somebody else the same problem (and maybe SPR'D it or is there any point in doing it when 4.5 is in distribution) Some background information of the tape: 1 The tape has many savesetts, about 7 2 Every savesett is about 5000 blocks and approx 600 files. The following is taken from crashdump (little reformatted). System crash information Time of system crash: 8-FEB-1987 18:35:30.52 Version of system: VAX/VMS VERSION V4.4 Reason for BUGCHECK exception: INVEXCEPTN, Exception while above ASTDEL or on interrupt stack Process currently executing: SAXEN Current image file: DUA0:[SYS0.][SYSEXE]BACKUP.EXE;1 Current IPL: 3 (decimal) General registers: R0 = 00000003 R1 = 00030000 R2 = 00000000 R3 = 7FF9091B R4 = 00000000 R5 = 00000000 R6 = 801D14D0 R7 = 7FFAD908 R8 = 7FF9FCCC R9 = 00000032 R10 = 7FF9FCC0 R11 = 00000072 AP = 7FF9FB4C FP = 7FF9FB30 SP = 7FF9FB00 PC = 80004862 PSL = 00030009 Processor registers: MicroVAX II P0BR = 802FF800 SBR = 004F7400 ASTLVL = 00000001 P0LR = 00000272 SLR = 00002080 SISR = 00000000 P1BR = 7FB0C800 PCBB = 001BD878 ICCS = 00000040 P1LR = 001FFAF1 SCBB = 004F5400 SID = 08000000 TODR = 23F7E77A SYSTYPE= 01010000 ISP = 80278C00 KSP = 7FF9FB00 ESP = 7FFE9E00 SSP = 7FFED04E USP = 7FF603C8 I'm very greatfull for any answers, and if someone wants to have more information from the crashdump i'll gladly send it. Thanks in advance!!! Hans Sundstroem E-mail earn/bitnet: SUNDSTRO@FINABO.BITNET Abo Akademi Heat Engineering Lab. Finland
cetron%utah-ced@UTAH-CS.ARPA.UUCP (02/09/87)
there was an early bug in the tk50's and vms where the invocation of backup with the norewind qualifier AND the verify qualifier would either hang for ever or crash the system.... This was especially irksome since the manual (in a dark obscure place) warned of not using /norewind/verify but neglected to mention that /verify was also no good since the /norewind was the default.... Try the same thing again without the verify, the bug might still be in there since it still was with 4.3... -ed cetron center for engineering design univ of utah
"Thomas_W._Taylor.WBST147"@XEROX.COM.UUCP (02/10/87)
Hans, Your problem may be due to the filename you are using for your saveset. While VMS4.x filenames may be 39.39 characters in length, the tape system doesn't like this. Try uisng 9.3 instead. Thomas W. Taylor Xerox Corp. Webster, N.Y. 14580