--------------------------------------------------------- NW4SAA.DOC -- 19961126 -- Email thread on NetWare for SAA --------------------------------------------------------- Feel free to add or edit this document and then email it back to faq@jelyon.com Date: Tue, 16 Apr 1996 15:26:17 EDT From: "Michael J. Dunigan" Subject: Re: Netware for SAA and HostPrint -Reply >We now have a Novell 4.1 Server running NWSAA v2.0 and >Host Print v1.1. Everything works great except if we ever have to >cycle the server, the statement at the bottom of AUTOEXEC.NCF > load hostprt is ignored. We have to manually type in the >command: > load hostprt and it works great. Can I add this command some >place such that we don't have to remember to go back to the server a >few minutes after starting the boot process and enter > load hostprt manually? Best I can tell is maybe we can't load >host print unless we have a valid connection to the mainframe at that >time. >I've had similar problems in the past, and it usually has to do with >the loading order. But since the command in question appears at the >bottom of AUTOEXEC.NCF, it's evidently something else, as you mention >below. Perhaps you just need a pause before the LOAD HOSTPRT >command. There's not a pause command for NCF files, unfortunately, >but there's one trick I've picked up. Put a plain, simple SET >command in before your load statement. Not SET SOMETHING=VALUE, just >SET. This will pause things at the console screen, showing you the >set categories, until you press the escape key. Hope that helps. I have a NW SAA 1.3b gateway with Hostprint 1.1 and I have found that you do need a pause before the LOAD HOSTPRT command. If the delay is not there, Host Print tries to load, and it has no sessions to "take over" because the profile has not yet completed loading. My solution to this was to get DELAYCMD from one of the mirror sites and use it to load the profiles and Host Print. With DELAYCMD.NLM, you can specify a time delay for the execution of the command submitted as a parameter on the LOAD DELAY command line. (It's easier to use than to explain.) Listed below is my .NCF file that loads the gateway: ; CSLoad.NCF ; Load the Communications Executive load commexec load dlm load sma ; Load NetWare for SAA load delaycmd 10 "csload nv_nodes" ; Load HostPrint for NetWare for SAA load delaycmd 30 "load hostprt" Another .NLM that you might want to get is called NCL.NLM. It adds some additional commands to the console processor, one of which is a Unix like AT command. I use this to automatically unload and reload the SAA gateway before and after my backup software runs, to make sure that all of the gateway files are closed when the backup software is running. I believe that NCL.NLM was written by Novell, and released as unsupported software. I have never had a problem with either DELAYCMD or NCL on my v3.11 server. (NW v4.1 comming soon, we'll see what happens then.) ------------------------------ Date: Tue, 26 Nov 1996 12:36:05 EST From: "Robert L. Herron" To: netw4-l@bgu.edu Subject: Re: Error -spurious hardware interrupt- and Netware 4.1 >About 4 months ago, I upgraded our AS400 gateway server from >NWSAA 1.3b to 2.0. Since it included a license to 3.12 or 4.1, I chose to >also upgrade the server to 4.1 (from 3.12). We had upgraded our main >file/print server to 4.1 and it seemed to make sense. I must have done >something wrong somewhere. If I log in to the SAA server and give the >wrong password (or have the wrong context), I get a message at the >console that says "spurious hardware interrupt 9 detected" Thinking that >the message may have something to do with the NIC, I replaced it. Not >once, but two times. Yesterday, when I knocked a cable from a workstation >on the same ring as the SAA server, it gave a message "cable has been >detached at the mau". These messages wouldn't be so bad, except the >workstations using the gateway to access the AS400 all drop their >connections. The only way to get them back is to restart the server. In >the meantime, the main file server blissfully continues servicing the >workstations, not noticing anything amiss. There is 20Mb of memory on the >SAA server (4 more than the formula asks for - small hard drive, few >AS400 sessions). Has anyone seen anything like this? Thanks in advance >for any help. I user NetWare for SAA 2 to connect to our AS/400 Advanced Series Model 510 (new PowerPC/RISC based system) with no problems. It runs on a Dell PowerEdge EL575 (Pentium 75, 32MB RAM, 1GB disk). Prior to SAA2, we use NetWare for SAA 1.3b on an old Northgate Elegance 325 (386DX-25, 16MB, 600MB disk). At one point this machine had 4 NIC cards installed (1 NIC per network segment (3) and the fourth NIC connected to the AS/400). The NICs were setup to user interrupts 9,10,11, and 15. When I tried running a shared folder to update PC Support, we got a spurious interrupt error. After many card changes, the errors would continue. While taking one of my CNE classes, I learned that NetWare did not like NIC cards using interrupt on the secondary IRQ controller (9-15). We moved the NICs to IRQs within the first controller and the problem went away. In short, I would try changing the NIC's IRQ to something lower. Also, make sure you've downloaded and applied SAA30A, SAA30B, SAA30C, and SAA30D.EXE from NetWire/Support Connection. These patches fix may little problems and make NWSAA much easier to work with. --------- Date: Tue, 26 Nov 1996 11:52:06 -0600 From: Richard French To: netw4-l@bgu.edu Subject: Re: Error -spurious hardware interrupt- and Netware 4.1 -Reply If you are using the IBM Token-ring driver, moving to version 3.24 or newer will fix this problem. --------- Date: Tue, 26 Nov 1996 13:07:06 -0500 (EST) From: Dennis Large To: netw4-l@bgu.edu Subject: Re: Error -spurious hardware interrupt- and Netware 4.1 >In short, I would try changing the NIC's IRQ to something lower. Also, >make sure you've downloaded and applied SAA30A, SAA30B, SAA30C, and >SAA30D.EXE from NetWire/Support Connection. These patches fix may little >problems and make NWSAA much easier to work with. AND SAA31. --------- Date: Wed, 27 Nov 96 00:17:41 -0800 From: Randy Grein To: "NetWare 4 list" Subject: Re: Error -spurious hardware interrupt- and Netware 4.1 >While taking one of my CNE classes, I learned that NetWare did not like >NIC cards using interrupt on the secondary IRQ controller (9-15). >We moved the NICs to IRQs within the first controller and the problem >went away. Hmm, first time I've seen that solution. I'm a little skeptical of the explanation, as I ALWAYS put ISA network cards at higher addresses, and have hundreds of happy servers out there. I have seen the problem over the years, and what it means is that Netware and the motherboard are having an argument over interrupts. This happens on a few machines frequently; the old Everex machines did it frequently. You could try moving the IRQ to 7 or below, or replace the motherboard with another brand. ------------------------------