OWEN@AUDUCVAX (Larry Owen) (02/03/90)
Below is the relevant portion of LMD.LOG. BN_SLAVE dies badly here, causing the old (Joiner-supplied) local mail delivery daemon to run (and junking up everyone's mailboxes with a bunch of un-replyable addresses). However, when jnet was restarted, BN_SLAVE ran just fine! Does anyone know what could cause this? -------------------------- LMD.LOG ---------------------- >$ ! LMD.COM - Command file for the JNET Local Mail Delivery Demon >$ ! This version invokes the PMDF server BN_SLAVE instead of the server >$ ! provided with JNET. This version is for JNET 3.2 or 3.3. >$ ! >$ set noon >$ define jan_sys jan_spool:,jan_root:[sys] >$ set default jan_sys: >$ if f$search("JAN_SYS:FANOUT.RSC") .nes. "" .and. - > f$search("JAN_SYS:FANOUT.PAR") .eqs. "" then create jan_sys:fanout.par >$ if f$search("JAN_SYS:FANOUT.PAR") .nes. "" then run - > /noaccounting /process_name="Fanout_Daemon" /error=jan_sys:fanout.err - > /output=jan_sys:fanout.err jan_sys:fno ! /priority=n >%RUN-S-PROC_ID, identification of created process is 00000119 >$ run /noaccounting /process_name="Profs_Daemon" /error=jan_sys:profs.err - > /output=jan_sys:profs.err jan_sys:profsin ! /priority=n >%RUN-S-PROC_ID, identification of created process is 0000011A >$ ! >$ if f$search("PMDF_ROOT:[EXE]BN_SLAVE.EXE") .eqs. "" then goto old_server >$ define/user pmdf_channel "bit_local" >$ define/user sys$output pmdf_root:[log]bn_slave.log >$ run pmdf_root:[exe]bn_slave.exe >%PAS-F-VARINDVAL, varying index value is greater than current length >%TRACE-F-TRACEBACK, symbolic stack dump follows >module name routine name line rel PC abs PC >BN_SLAVE STRIP 1101 0000005C 00004EEC >BN_SLAVE MUNGED 1148 0000019F 00003E8B >BN_SLAVE BN_SLAVE 1323 000005B4 000047BC >$ ! >$ old_server: >$ define jan_mailprot "BITNET%" >$ define sys$scratch jan_sys: etc. Thanks in advance for any insight. Larry Owen Mgr., Network Support Auburn University Bitnet: owen@auducvax Internet: owen@ducvax.auburn.edu
NED@HMCVAX.CLAREMONT.EDU (Ned Freed, Postmaster) (02/04/90)
Larry Owen writes: > Below is the relevant portion of LMD.LOG. BN_SLAVE dies badly here, > causing the old (Joiner-supplied) local mail delivery daemon to run > (and junking up everyone's mailboxes with a bunch of un-replyable > addresses). However, when jnet was restarted, BN_SLAVE ran just fine! > Does anyone know what could cause this? > ... >%PAS-F-VARINDVAL, varying index value is greater than current length >%TRACE-F-TRACEBACK, symbolic stack dump follows >module name routine name line rel PC abs PC >BN_SLAVE STRIP 1101 0000005C 00004EEC >BN_SLAVE MUNGED 1148 0000019F 00003E8B >BN_SLAVE BN_SLAVE 1323 000005B4 000047BC The fact that BN_SLAVE is dying in STRIP means that you're running a PMDF prior to 3.0-20. STRIP was eliminated entirely in that patch. It also probably means that you're running a PMDF prior to 3.0-6, since 3.0-6 fixed a problem like this specifically in STRIP. A number of problems in BN_MUNGE have been found at one time or another. They are, as a rule, triggered by some funniness in one or more header lines. Problem is, BN_SLAVE dies and JNET picks up and processes the message causing the problem, leaving no clue as to what specifically caused the trouble... You have three options: (1) Install 3.0-6 and hope it fixes it. (2) Install 3.0-20 (which totally supercedes 3.0-6) and eliminate this problem for sure. (3) Upgrade to 3.1, also eliminating the problem for sure. Hope this helps. Ned