Dear All
I have done a bit more testing. My original server is a P133 with 32MbRAM, I know its less than suggested, but it seemed to work initially, and then this printing problem showed up after a day or two.
This machine has problems when I print to both printers, and gives port unavailable errors and leaves print jobs in the queue. I can only print 2 jobs before the error messages start and then I cannot print anything, unless I empty the print queues on a Windows95/98 workstation.
I also added some spare pieces of RAM I had making it 48Mb but still had the same problems.
My test server is a P166 with 80Mb RAM and I only had it set up to print to one printer, I noticed that it would print OK as many print jobs as I wanted to, except that two print jobs were still left in the print queue all the time, even though they had been printed.
I connected a second parallel port card and second printer to the test server and both printers worked OK, with as many print jobs as I could send, except I still saw 2 print jobs in each printers queue, even though these had been printed.
I swapped network cards and also swapped parallel port cards, but still had the problem on the original server. I could not swap RAM as it is different size and shape, and some is buffered some is not.
I also put a different hard drive into the original server that had e-smith v4.1.2 on it, and both printers printed OK without any jobs remaining in the printer queues with only 32Mb RAM (same as for 5.1.2).
I also tried a P600 with 256Mb RAM & 5.1Gb HDD with the second parallel port card in it with a fresh instal of v5.6.
I could print to both printers OK, but still saw 2 print jobs in the printer queues even after they had been printed.
Conclusions
I would say that the RAM in my original server is not enough, thats what Mitel are saying anyway, but I have found a printing problem that partially disappears with 80Mb instead of 32 MB or 48Mb
The test server consistently prints with 80Mb RAM, so I must conclude more RAM is part of the answer.
BUT on all three of the servers I tried I still had the problem of print jobs staying in the queue and not being flushed, so it seems that there is something wrong with v5.6 or the kernel and/or rpms it uses. Unfortunately my knowledhge of these things is insufficient to be able to diagnose further.
Anybody have any comparisons of their systems configurations, or any comments/suggestions as to why the jobs are staying in the print queue.
Regards
Ray Mitchell