Re: printing problem with recent lp patch

Gary Mills wrote:
In <1246983030.55460@news1nwk> Martha Starkey <martha.starkey@xxxxxxx> writes:

On 06/26/09 04:41, Martin Paul wrote:
Martha Starkey wrote:
In the meantime, we found that many of these "can't print" problems with S10U5+ are because either one or both of the 'psm-lpd' and 'psm-ipp' print packages are missing. Within this discussion seems like a good place to list ALL the print packages for S10u5+:
Just FYI - I have all of them on my systems. Using the default "Entire Distribution" has been the cluster of choice for me for years; it saves you from obscure issues with patching as well.

The initial problem reported to us (to which I was comparing your issue to) turned out to be a combination of configuration problems and missing/incorrect-for-the-release print packages. We got it straightened out and the queue working.

The "fix for the fix" for CR6802265 will be CR6841806, which looks to be in S10 U8 and for which a patch will be forthcoming (standard disclaimers apply!).

I can't find either of those bug numbers in the three visible bug
reporting systems.

All the CRs are internal, meaning you won't see them. You may run across them in opensolaris.org changelogs, but just a listing and no details.

However, my truss of `lp /usr/pub/ascii' shows
that the problem lies with /usr/lib/print/lpd-port:

1593: open("/var/run/syslog_door", O_RDONLY) = 4
1593: door_info(4, 0x08046920) = 0
1593: getpid() = 1593 [1591]
1593: door_call(4, 0x08046958) = 0
1593: close(4) = 0
1593: ioctl(1, I_SENDFD, 4) Err#9 EBADF
1593: read(4, 0x0804756C, 1024) Err#9 EBADF
1593: fxstat(2, 3, 0x08046B28) = 0
1593: time() = 1247244764
1593: getpid() = 1593 [1591]
1593: putmsg(3, 0x080474E8, 0x080474F4, 0) = 0
1593: open("/var/run/syslog_door", O_RDONLY) = 4
1593: door_info(4, 0x08046940) = 0
1593: getpid() = 1593 [1591]
1593: door_call(4, 0x08046978) = 0
1593: close(4) = 0
1593: _exit(9)

When I replace that file with the copy that was delivered before I
applied patch 141779-02, printing works again. Without that change,
my output never reached the printer.

My hope, at the outset of this thread, was that there was an already known issue that could be easily identified. I've given up on that. I should know better, but at my core, I'm an optimist.

No doubt something is going on b/t 140397-06 and -08. I was pointed to 6802265 being the most likely culprit. I'm finding that with all the CRs being internal, the best advice would be to open a case with Support so a proper investigation can be done.

if you do so, you'll want to know: OS of print server and clients (/etc/release and uname -a on the Solaris boxes). /etc/printers.conf (or an Explorer output). lpr-debug from a failed job from any Solaris machines in the mix. What was the lp patch level (if any) prior to the 'problem' patch. Full truss outputs and a snoop would also be helpful.



.

Relevant Pages

  • Re: printing problem with recent lp patch
    ... with S10U5+ are because either one or both of the 'psm-lpd' and 'psm-ipp' print packages are missing. ... Within this discussion seems like a good place to list ALL the print packages for S10u5+: ... The "fix for the fix" for CR6802265 will be CR6841806, which looks to be in S10 U8 and for which a patch will be forthcoming. ...
    (comp.unix.solaris)
  • Re: printing problem with recent lp patch
    ... 'psm-ipp' print packages are missing. ... like a good place to list ALL the print packages for S10u5+: ... I can't find either of those bug numbers in the three visible bug ... reporting systems. ...
    (comp.unix.solaris)
  • Re: printing problem with recent lp patch
    ... S10U5+ are because either one or both of the 'psm-lpd' and 'psm-ipp' print packages are missing. ...
    (comp.unix.solaris)