tim@cs.fau.edu (Tim McKenzie) (04/04/91)
My prsvr keeps aborting from when called within the /etc/rc.user file. When I run the same command with the same config file all goes great. I have tried cpo, cps, and /bin/nohup in front of the command below, but with no success in keeping the sucker alive from bootup. /sys/hardcopy/prsvr /sys/hardcopy/hp_config & The prmgr and prsvr are both on the same machine. and the prmgr stays alive. HP/Apollo doesn't seem to be much help. It's running on a DN-4500 with Domain OS 10.2 Thanks in advance. ---------------------------------------------------------- | Tim McKenzie | "I called the witchdoctor and he | | tim@cs.fau.edu | told me what to do." | | 131.91.80.13 | --David Seville | ----------------------------------------------------------
rpj@echo.canberra.edu.au (Ross Johnson) (04/08/91)
In <1991Apr04.001427.13506@cs.fau.edu> tim@cs.fau.edu (Tim McKenzie) writes: >My prsvr keeps aborting from when called within the /etc/rc.user >file. When I run the same command with the same config file >all goes great. I have tried cpo, cps, and /bin/nohup in front >of the command below, but with no success in keeping the sucker >alive from bootup. Seems to be a permissions problem if it runs as "root" but not as "user.server". Check the permissions on your /sys/mgrs/print_* files, they should be mode 755 with owner="user", group="server". -- +----------------------+---+ | Ross Johnson | | E-Mail: rpj@ise.canberra.edu.au | Info Sciences and Eng|___| | University of Canberra | UUCP : uunet!munnari!ise.canberra.edu.au!rpj | PO Box 1 | JANET : rpj%au.edu.canberra.ise@EAN-RELAY | Belconnen ACT 2616 | BITNET: rpj%ise.canberra.edu.au@relay.cs.net | AUSTRALIA | +--------------------------+
obrennan@CC3.CC.UMR.EDU (obrennan) (04/08/91)
My prsvr keeps aborting from when called within the /etc/rc.user file. When I run the same command with the same config file all goes great. I have tried cpo, cps, and /bin/nohup in front of the command below, but with no success in keeping the sucker alive from bootup. /sys/hardcopy/prsvr /sys/hardcopy/hp_config & The prmgr and prsvr are both on the same machine. and the prmgr stays alive. HP/Apollo doesn't seem to be much help. It's running on a DN-4500 with Domain OS 10.2 Thanks in advance. ---------------------------------------------------------- | Tim McKenzie | "I called the witchdoctor and he | | tim@cs.fau.edu | told me what to do." | | 131.91.80.13 | --David Seville | ---------------------------------------------------------- ------------------------------------------------------------------------------------------- Two things I can think of: 1) the process that you are running in the background from the /etc/rc.user file will run as user.server. What we have found is that you need to make group server an owner for all files in /sys/hardcopy/drivers and /sys/mgrs before the print server will run under user.server. Don't forget to change the initial file acls in the drivers and mgrs directories. 2) starting a print server in the RC.USER file has never seemed to be consistent since SR10.2. My best guess is that the print server starts before the line is initialized (or something of the sort) and the initialization aborts the prsvr. Just a guess though. Gerry O'Brennan Programmer/Analyst II Computing Services University of Missouri - Rolla ------------------------------ obrennan@apollo.cc.umr.edu c0022@umrvmb.umr.edu ------------------------------
dbfunk@ICAEN.UIOWA.EDU (David B Funk) (04/09/91)
in posting <1991Apr04.001427.13506@cs.fau.edu>, Tim McKenzie writes: >My prsvr keeps aborting from when called within the /etc/rc.user >file. When I run the same command with the same config file >all goes great. I have tried cpo, cps, and /bin/nohup in front >of the command below, but with no success in keeping the sucker >alive from bootup. >/sys/hardcopy/prsvr /sys/hardcopy/hp_config & This sounds like either an ACL problem or some kind of timing problem. The daemons that are started from the rc.user file run with the SID of "user.server.none". Have you tried creating an account "user.server.none", logging in to it and then running the prsvr from that shell? This is one sure-fire way to find ACL problems. Under sr10.2, the prsvr wants to create a manager object in "/sys/mgrs" so if the world (or user.server.none) doesn't have write rights to /sys/mgrs, then your prsvr won't run from the rc.user file. It is possible that the prsvr is getting started before the prmgr is ready to register it. You might try to move the prmgr to a different node, so that it is not booting at the same time that the prsvr is. Check the system proc_dump log to see if prsvr is leaving a trace-back error report. Just after boot up do a "tb -last 3" to see what the last few process crash reports were. Also you could try adding a std-out & err-out redirection in the "rc.user" file to direct any error messages from prsvr into a file. Try something in your rc.user like: (/sys/hardcopy/prsvr /sys/hardcopy/hp_config > /tmp/prsvr.log 2>&1 ) & This will direct all messages from prsvr into the file "/tmp/prsvr.log" You don't want to run prsvr this way for normal operation, as prsvr is rather "chatty" and that file would grow with continued use. However it is good for a debugging session. Dave Funk