5.1b-1 tcp printing problems.

[peter carpenter_1]

I currently have a support case open with HP on this, but thought I would post the details in case anyone has experienced similar problems.

We recently upgraded our 5.1a clustered live system which consists of a GS160 and an ES45 to 5.1b-1 (NHD7 + PK3). We then applied the patch that corrects the premature tcp disconnection problem. Several hours after completing and handing back to our users, we started to get reports of jobs being submitted to printers and only the first 1 or 2 jobs printing completely and then stopping. Examining the queue, we can see the remaining jobs waiting, and the message 'attempting to print...'. Our printers are mostly of the Tally range, and are attached to DEC 90TL terminal servers. We have about 250 of these company wide. netstat showed several time_wait connections to these.

We do a fairly large volume of printing everyday, with jobs ranging from 1 page to 80.

We set lpr.log into debug, but did not see anything obvious in there. An lpc restart appears to kick the daemon back into life, but only for the next 1 or 2 jobs and then it hangs again. We use the jj#1 parameter to speed things up, and doing a ps -ef |grep lpd shows the spawned lpd processes. All jobs that print out are printed out completely, ie we havent seen a printer give up half way through a print job like the premature tcp disconnection problem.

The strange thing is, we upgraded our test cluster (2 ds10s) to 5.1b-1 with the tcp patch and did a fairly large print test which ran sucecssfully. We have not yet been able to re-create the problem on our test platform, but will involve all our remote sites in a large scale test. HP have also not been able to re-produce the problem I'm describing, but are still investigating. For the time being we have had to roll the live cluster back to 5.1a and the problem has gone away.

I am just waiting for some feedback from HP as to our next step, but will post the details here if anyone is interested.

[Ralf Puchner]

Peter,

try the following:
remove OF= statement from printcap (known problem). If this will help install patchkit #4 for 5.1B containing a fix (QAR97891).

[peter carpenter_1]

Many thanks for your post, we are due to arrange a time to try and re-create the problem in the test environment, so I will bear this information in mind if the problem re-emerges. Will let you know how we get on.

[Michael Schulte zur Sur]

Ralf,

we have a similar problem. The print of large jobs does not finish. Is that caused by the same problem?
We also have 5.1B-1 pk3.
If we drop of then how can we create the cr necessary?
Is there single patch, which can be installed without rebooting?

thanks,

Michael

[Ralf Puchner]

Michael,

please open a call within the support center, guessing without analyzing the problem is not an issue...