weis@netmbx.UUCP (Dietmar Weis) (04/24/91)
On a DECSystem3100 running Ultrix 4.0 sometimes occur inconsistencies in the lpd system (printing to lat printers), followed by a system crash!! lpc status said for example "no progress since 3 days" , "waiting to connect" or "no daemon active", the error log says "output filter to lat malfunctioned (0)", "cannot open /dev/tty??, no such file or directory" or "job aborted". But there *was* a lpd daemon, there *is* a /dev/tty??, but there were also an idle /usr/lib/lpdfilters/xf with PPID 1 and the forked /bin/cat-process. The user did not miss his last printouts. So, cleaning up the queue, restarting and killing the idle print-processes (I can't remember in which order, doesn't matter) resulted in an immediate system crash: cpu0 panic: unaligned access locks held by cpu0 print locks held by non-active processes Rebooted to Multiusermode the system crashed again and again, sometimes at the starting up of lpd in rc.local, sometimes after login and doing lpc status. At this time there were of course no more print processes and no print jobs. Only powering off the terminal server resolved this. This is the 6. time something like this happend and it's very serious. Other messages in the lpd error log were: "sleeping, retrying in xx seconds", (already discussed here, with no solution I think) and "socket already in use". So, what is going on here, does anyone know? Specification: DS3100 with 20 MB, Ultrix 4.0, Emulex Terminal Server P4000, Kyocera F-800 printers. Thank you very much, Dietmar
tamura@versatc.versatec.COM (Mark Tamura) (04/27/91)
In article <2286@netmbx.UUCP>, weis@netmbx.UUCP (Dietmar Weis) writes: > > On a DECSystem3100 running Ultrix 4.0 sometimes occur inconsistencies > in the lpd system (printing to lat printers), followed by a system crash!! > > lpc status said for example "no progress since 3 days" , > "waiting to connect" or "no daemon active", the error log says "output > filter to lat malfunctioned (0)", "cannot open /dev/tty??, no such file or > directory" or "job aborted". I had the same problem. Apparent with Ultrix 4.0 the system will crash if no output filter is specified in /etc/printcap. After I added an "of" entry, everything worked fine. Also Digital support refused to make/send a patch for a third-party printer. -- Xerox Engineering Systems INET: Mark_S_Tamura.SCB1Versatec@Xerox.COM West Coast Operations UUCP: {ames,pyramid,sun}!versatc!tamura Versatec Products AT&T: (408)988-2800 x5202
thomas@mipsbx.lkg.dec.com (Matt Thomas) (04/28/91)
In article <21639@versatc.versatec.COM> tamura@versatc.versatec.COM (Mark Tamura) writes: >In article <2286@netmbx.UUCP>, weis@netmbx.UUCP (Dietmar Weis) writes: >> >> On a DECSystem3100 running Ultrix 4.0 sometimes occur inconsistencies >> in the lpd system (printing to lat printers), followed by a system crash!! >> >I had the same problem. Apparent with Ultrix 4.0 the system will crash if >no output filter is specified in /etc/printcap. After I added an "of" entry, >everything worked fine. Also Digital support refused to make/send a patch >for a third-party printer. A patch for the CSCs has been available for V4.0 for a long time (the bug was detected right after V4.0 was shipped) and was incorporated into V4.1. As for the CSC not sending the patch, my opinion is that they should have sent the patch regardless of the printer. The bug is independent of the type of printer and type of DECserver. Sigh... Matt Thomas Internet: thomas@pa.dec.com DECnet-ULTRIX Development UUCP: ...!decwrl!thomas Digital Equipment Corporation Disclaimer: This message reflects my own Littleton, MA warped views, etc.