jmg@cernvax.UUCP (john gerard) (11/24/89)
We are trying to run a LAT service through a DS3100 on a lot of ttys. We have configured 64 ttys with MAKEDEV, and we use ttys from tty04 for a lat/telnet service (actually a lat/tn3270 service). So far, so good. If we configure too many of these ttys in the /etc/ttys file then at some time the system goes bananas. Shells get continuously created and destroyed (the lat/telnet service program calls in a shell script), zombies abound, the CPU is 100% busy etc. etc. Currently all is fine with ttys from tty04 to tty40 inclusive. However, if I add another 10 or so in /etc/ttys then the bananas reappear. Maybe it is the configuration, but I don't think so: we reconfigured with higher limits today but the problem is just the same. I recall that some time ago someone talked about limitations on the number of ttys per each multiple LAT service. This was referring (I think) to Ultrix 3.0: we have now Ultrix 3.1 and also a set of patches (not yet applied, because we are not sure whether they would help or hinder) but we cannot see any relevant information. Also, the previous posting (from George Robbins) did not specify a numerical limit. Has anyone an update, or (even better) a fix for this problem? Mike Gerard ps. Does anyone else note that when a DECterm VT300 window starts up with a host status line enabled this line cannot be used? We have to disable it then re-enable it. Applies also to VMS DECwindows. -- _ _ o | __ | jmg@cernvax.uucp | | | | _ / \ _ __ _ __ _| jmg@cernvax.bitnet | | | | |_) /_) | __/_) | (___\ | (_/ | J. M. Gerard, Div. DD, CERN, | | |_|_| \_/\___ \__/ \___| (_|_| \_|_ 1211 Geneva 23, Switzerland
neil@cpd.com (Neil Gorsuch) (11/25/89)
In article <1152@cernvax.UUCP> jmg@cernvax.UUCP (john gerard) writes: >We are trying to run a LAT service through a DS3100 on a lot of ttys. >We have configured 64 ttys with MAKEDEV, and we use ttys from tty04 >for a lat/telnet service (actually a lat/tn3270 service). So far, so good. >If we configure too many of these ttys in the /etc/ttys file then at >some time the system goes bananas. Shells get continuously created and >destroyed (the lat/telnet service program calls in a shell script), >zombies abound, the CPU is 100% busy etc. etc. >I recall that some time ago someone talked about limitations on the number >of ttys per each multiple LAT service. If the problem is something in the LAT software, a possible solution is to use our SCSI serial/parallel servers. From what I have been told of LATs (we don't have any to compare against), they have higher performance. And they certainly cost a lot less than LATs and relieve a lot of ethernet load. Plus you get a Centronics parallel port for each 8 serial ports. -- Neil Gorsuch INTERNET: neil@cpd.com UUCP: uunet!zardoz!neil MAIL: 1209 E. Warner, Santa Ana, CA, USA, 92705 PHONE: +1 714 546 1100 Uninet, a division of Custom Product Design, Inc. FAX: +1 714 546 3726 AKA: root, security-request, uuasc-request, postmaster, usenet, news
jmg@cernvax.UUCP (john gerard) (11/27/89)
In article <1989Nov25.015814.26839@cpd.com> neil@uninet.UUCP (Neil Gorsuch) writes: >>We are trying to run a LAT service through a DS3100 on a lot of ttys. >>We have configured 64 ttys with MAKEDEV, and we use ttys from tty04 >>for a lat/telnet service (actually a lat/tn3270 service). So far, so good. >>If we configure too many of these ttys in the /etc/ttys file then at >>some time the system goes bananas. Shells get continuously created and >>destroyed (the lat/telnet service program calls in a shell script), >>zombies abound, the CPU is 100% busy etc. etc. >>I recall that some time ago someone talked about limitations on the number >>of ttys per each multiple LAT service. > >If the problem is something in the LAT software, a possible solution >is to use our SCSI serial/parallel servers. From what I have been >told of LATs (we don't have any to compare against), they have higher >performance. And they certainly cost a lot less than LATs and relieve >a lot of ethernet load. Plus you get a Centronics parallel port for >each 8 serial ports. Unless I have misunderstood, this solution sounds like asking me to buy some extra equipment in order to get around a software bug. However, I don't really understand it. We have lots of people who already have their terminals on DECservers and who want one of their sessions to go to an IBM. This is currently done by having a bank of reverse DECservers with lines going into an IBM 7171 so as to get the 3270-VTxxx conversion. I leave you to calculate the line cost of these reverse DECservers plus cables plus 7171 ports! Where does this SCSI connection fit in? What I want is to allow them to have a session through the DECstation, which will do the 3270-VTxxx itself. I can foresee a use for some kind of multiplexor to allow real physical lines to hang off a DECstation (via SCSI I suppose), and would be happy to have information on the "best" way to do this. However, it does not solve my current problem. Have I written a long enough follow-up such that it gets accepted? I sure hope so. -- _ _ o | __ | jmg@cernvax.uucp | | | | _ / \ _ __ _ __ _| jmg@cernvax.bitnet | | | | |_) /_) | __/_) | (___\ | (_/ | J. M. Gerard, Div. DD, CERN, | | |_|_| \_/\___ \__/ \___| (_|_| \_|_ 1211 Geneva 23, Switzerland
jmg@cernvax.UUCP (john gerard) (12/11/89)
Mea culpa. There may be some restriction on the number of ttys per LAT service, but that was not my problem (which was an error in setting up the lta devices with MAKEDEV). -- _ _ o | __ | jmg@cernvax.uucp | | | | _ / \ _ __ _ __ _| jmg@cernvax.bitnet | | | | |_) /_) | __/_) | (___\ | (_/ | J. M. Gerard, Div. DD, CERN, | | |_|_| \_/\___ \__/ \___| (_|_| \_|_ 1211 Geneva 23, Switzerland