Thanks to the lone response from Phil Lorgan. My original query related
to an apparent problem with a PI/open RBMS hanging when using lpr.
We have now identified an issue with the BSD spooler, where lpd children
die before printing. PI/open waits for the print job to finish before it
submits another (as it maintains its own expanded spooler environment),
but as the lpd process has died, the job never prints. Hence the hang!
Things are not so bleak at the SunOS level, as lpr causes a new lpd child,
which then prints the "blocked" job. The tech guys at Sun, when we showed
them the problem, said "Ahhh, that's why customer X is having that hassle"
(which was gratifying in one sense, but annoying in another). This
lpr/lpd interaction also accounts for why nobody has reported it as a
SunOS bug (has anyone else seen this? What did you do, ignore it or shrug
and utter a few choice words?).
The fix is to write a simple script to check for the lpq response "no
daemon present" and use lpc restart. We've done that and all is OK. Sun
will probably not fix the bug as they are focused on Solaris 2.
Now, my only other query, and it is FYI only, is: was the BSD spooler
written so "no daemon present" is a common and expected occurrence?
Reading the manual pages of lpc (see the "restart" command) indicates
this, but why??? Any BSD guru's care to comment :)
+---------------------------*-------------------------------------+
| Michael Panosh | Internet: mwp.michael@melpn1.cv.com |
| Network Applications | |
| Computervision Pty Ltd | |
| 6/10 Queens Road | Ph: (+613) 869 5777 |
| Melbourne, 3004 | Fax: (+613) 867 5306 |
| Australia | |
+---------------------------*-------------------------------------+
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:07:52 CDT