jww@bonnie.UUCP (Joel West) (05/21/85)
I am having problems with my uucp. I don't have source, so I'll probably end up bugging the vendor if it's not something I can do via configuration change. Symptoms: This involves two successive polls to any arbitrary site. I do a 2% /usr/lib/uucp/uucico -r1 -sbonnie & and it works just fine. Then, after the process terminates and the LOGFILE shows "conversation complete", if I do a 3% /usr/lib/uucp/uucico -r1 -ssdcc3 & it dies, leaving a LCK..culoo and a LCK..sdcc3 If I remove the LCK files and do a 4% /usr/lib/uucp/uucico -r1 -sanywhere & it works perfectly. This good/bad alternation is completely repeatable, does not depend on the site and seems to depend entirely on whether this is the "odd" or "even" uucico request. There is some evidence this applies to odd/even tip (cu) requests but I can't prove it. The failed job dies with an unexpected SIGALRM signal (under BSD job control it prints "Alarm Clock" on the termination message). I assume that alarm(2) is used for a timeout, but for some reason, in the "odd" cases this signal comes through untrapped and thus kills the jobs with the LCK.. files intact. Using the "-x9" seems to indicate that the connection is being processed when it dies; the only LOGFILE entry is "LOCKED". Configuration: Gould UTX-32 v1.1 (ihnp4!bonnie!gould9!joel) BSD port to Gould hardware includes BSD job control, virtual kernel Some System V features, including the commands: uulog uux Any ideas would be most appreciated. -- Joel West (619) 457-9681 CACI, Inc. - Federal 3344 N. Torrey Pines Ct La Jolla 92037 jww@bonnie.UUCP (ihnp4!bonnie!jww) westjw@nosc.ARPA "The best is the enemy of the good" - A. Mullarney