dennis@nosc.mil (Dennis Cottel) (05/13/91)
Anyone care to summarize the high (and low) points of the ADUS SysAdmin conference for those of us who couldn't attend? Dennis Cottel, dennis@NOSC.MIL, (619) 553-1645 Naval Ocean Systems Center, San Diego, CA 92152
rfh3273@galileo.rtn.ca.boeing.com (Dick Harrigill) (05/25/91)
In article <dennis.674146246@woodstock> dennis@nosc.mil (Dennis Cottel) writes: >Anyone care to summarize the high (and low) points of the ADUS SysAdmin >conference for those of us who couldn't attend? I'll let someone else speak to the high points. The low point was absolutely a presentation of a non-Apollo non-Domain box called the 700. This is a fine box, but it is not Domain and therefore should not have been presented at the ADUS (Apollo Domain User Society) meeting. They touted a fantastic price/performance ratio. However, to Apollo Users the price/performance is infinity. Another low was the announcement by HP of a proposal to nodelock the OS! I sure hope they abandoned that one after the sparks that flew. -- Dick Harrigill, an independent voice from: Boeing Commercial Airplanes M/S 9R-49 PO BOX 3707 Renton Avionics/Flight Systems Seattle, WA 91824 Computing Support (206) 393-9539 rfh3273@galileo.rtn.ca.boeing.com
mroussel@alchemy.chem.utoronto.ca (Marc Roussel) (05/28/91)
In article <399@galileo.rtn.ca.boeing.com> rfh3273@galileo.rtn.ca.boeing.com (Dick Harrigill) writes: >Another low was the announcement by HP of a proposal to nodelock the OS! Could someone explain what this means? Marc R. Roussel mroussel@alchemy.chem.utoronto.ca
chen@digital.sps.mot.com (Jinfu Chen) (05/29/91)
> >Anyone care to summarize the high (and low) points of the ADUS SysAdmin > >conference for those of us who couldn't attend? > >I'll let someone else speak to the high points. I didn't attend the conference but a fellow worker passed a few (high?) points: In SR10.4 (scheduled released on Q3/91?) disk quota is enforced in single volume (i.e. can't put a user's files across several volume). Also in SR10.4, screen lock will be provided for DM (I guess it will use the value set by scrto). OSF DCE will run on ATR (Apollo Token Ring) cable. At least we don't have to re-wire the building with one exception (see below) > They touted a fantastic price/performance ratio. However, to >Apollo Users the price/performance is infinity. Exactly! Software issues aside, HP doesn't even offer an ATR option on the 700 series so for those of us who spent zillions dollars (or hours if your boss is cheap) setting token ring cables, we couldn't hook up the 700 series into our existing network even OSF/1 and DCE would have been available. An unconfirmed rumor is that HP contracted someone else to make a ATR board for the 700 series. I hope this is true and HP will make it available in similar fasion on the 400 series. Like many have said before, the 700 series bears nothing of Apolloism except the label. To current Domain users, it's just another fast box (I'm impressed with the speed though).
rees@pisa.citi.umich.edu (Jim Rees) (05/29/91)
In article <51d64a3b.3593b@digital.sps.mot.com>, chen@digital.sps.mot.com (Jinfu Chen) writes:
Exactly! Software issues aside, HP doesn't even offer an ATR option on the
700 series so for those of us who spent zillions dollars (or hours if your
boss is cheap) setting token ring cables, we couldn't hook up the 700 series
into our existing network even OSF/1 and DCE would have been available.
Just curious -- has anyone tried running ethernet over existing token ring
cable? If you set it up as thinnet sections, and keep them short, it just
might work. Of course your taps will be at the wrong places, but I've been
amazed at what has worked around here in the past.
My home Apollo ring is currently wired up with thinnet cable, and that works
fine.rfh3273@galileo.rtn.ca.boeing.com (Dick Harrigill) (05/30/91)
In article <1991May27.203417.15155@alchemy.chem.utoronto.ca> mroussel@alchemy.chem.utoronto.ca (Marc Roussel) writes: >>In article <399@galileo.rtn.ca.boeing.com> rfh3273@galileo.rtn.ca.boeing.com >>(Dick Harrigill) writes: >>Another low was the announcement by HP of a proposal to nodelock the OS! > > Could someone explain what this means? HP's Robert Clawson, who heads up the Domain/OS effort in Chelmsford, gave a pitch entitled, "HP Workstation Operating System Update." The main focus was the "future" of Domain/OS and OSF/1. He mentioned the fact that with each copy of an OS, such as SysV, a royalty must be paid. Since the introduction of OSF/1 will further muddy the waters of availability, it is conceivable that a user would purchase/license only one UNIX flavor, but still run the others and avoid the royalties. Thus, they are looking at node locking the OS. This comment was not received well and seemed to destroy an otherwise informative presentation. -- Dick Harrigill, an independent voice from: Boeing Commercial Airplanes M/S 9R-49 PO BOX 3707 Renton Avionics/Flight Systems Seattle, WA 91824 Computing Support (206) 393-9539 rfh3273@galileo.rtn.ca.boeing.com
wjw@ebh.eb.ele.tue.nl (Willem Jan Withagen) (05/30/91)
In article <51d75679.1bc5b@pisa.citi.umich.edu>, rees@pisa.citi.umich.edu (Jim Rees) writes: => In article <51d64a3b.3593b@digital.sps.mot.com>, chen@digital.sps.mot.com (Jinfu Chen) writes: => => Exactly! Software issues aside, HP doesn't even offer an ATR option on the => 700 series so for those of us who spent zillions dollars (or hours if your => boss is cheap) setting token ring cables, we couldn't hook up the 700 series => into our existing network even OSF/1 and DCE would have been available. => => Just curious -- has anyone tried running ethernet over existing token ring => cable? If you set it up as thinnet sections, and keep them short, it just => might work. Of course your taps will be at the wrong places, but I've been => amazed at what has worked around here in the past. => => My home Apollo ring is currently wired up with thinnet cable, and that works => fine. My 400dl doesn't have an ring-control, nor can it ever have it. Hence I'm using thinnet. We first tried to use one of the 92ohm arcnet cables, but the tried already implies the negative success. The we have this very thin, but still 50ohm coax in the wall. Nobody use it so we started using that. Currently my segment is 65m long, and in my room are 1 400dl and 2 PC-wd8003E card connected. This works, but measuring the lines tells that the ac and dc-resistance is already to big. (~>0.01 ohm) Trying to append another segment of 65m of this cable collapses the system. My apollo won't boot, the controller says that the controller has an error. An looking at the signals, things are really messy. So maybe thinnet can be used for a ring (What's the impedance of ring-cable) But a regular ethernet is much more nasty. Ciao, Willem Jan -- Eindhoven University of Technology DomainName: wjw@eb.ele.tue.nl Digital Systems Group, Room EH 10.10 P.O. 513 Tel: +31-40-473401 5600 MB Eindhoven The Netherlands
rees@pisa.citi.umich.edu (Jim Rees) (05/30/91)
In article <1198@eba.eb.ele.tue.nl>, wjw@ebh.eb.ele.tue.nl (Willem Jan Withagen) writes:
So maybe thinnet can be used for a ring (What's the impedance of ring-cable)
But a regular ethernet is much more nasty.
I thought ethernet and ring were both 75 ohm, but I may be mis-remembering.
I don't have my Xerox blue book in front of me.dbfunk@ICAEN.UIOWA.EDU (David B Funk) (05/31/91)
IN posting <51d64a3b.3593b@digital.sps.mot.com>, Jinfu Chen writes: > OSF DCE will run on ATR (Apollo Token Ring) cable. At least we don't have to > re-wire the building with one exception (see below) > > > They touted a fantastic price/performance ratio. However, to > >Apollo Users the price/performance is infinity. > > Exactly! Software issues aside, HP doesn't even offer an ATR option on the > 700 series so for those of us who spent zillions dollars (or hours if your > boss is cheap) setting token ring cables, we couldn't hook up the 700 series > into our existing network even OSF/1 and DCE would have been available. > > An unconfirmed rumor is that HP contracted someone else to make a ATR board > for the 700 series. I hope this is true and HP will make it available in > similar fasion on the 400 series. Ok folks, rumor control time. ATR is not an option on the 700 series WHEN RUNNING HP-UX. HP-UX has no idea how to talk to any kind of token ring board. However ATR -IS- an option for the 700 series under OSF. HP/Apollo has said time & again that they -are- going to support OSF on ATR; even to the extent of running both Domain/OS & OSF machines on the same net. The 700 series hardware can run an ATR, it is just a matter of the OS software drivers. All the 700 series machines can take at least 1 EISA bus card. The 750 has a cage with 4 EISA slots, the 730 has 1 standard EISA slot, and the 720 has one optional EISA slot. On the 720 it is down by the power supply and if you don't order the optional adapter for it, then it is hidden by a cover plate. Given that HP/Apollo already has a source of ISA bus ATR boards I doubt that they need to go to somebody else for them. Maybe they are looking for a new source of IBM token ring boards (16 megabit maybe?). We are ordering ATR with our 750 ;). Our salesman did not have the part number for it in his current catalog but he assures us he will find it, even if he has to call Chelmsford. (no we are not buying it to run HP-UX ;). > Like many have said before, the 700 series bears nothing of Apolloism except > the label. To current Domain users, it's just another fast box (I'm impressed > with the speed though). Gee, I thought that I detected a strong familial resemblance between the 700 series and the 400t/s series; even to "das blinken lites" ;=). It is true that Domain/OS is not available for them but OSF/1.0 is due out, Real Soon Now. The 400 series is next on the OSF development list. Dave Funk PS: It is true that the new low cost 425e (woody) does not have ATR capablilty. This box was designed to be as cheap as possible and does not have any kind of slots at all. Everything is designed into the mother board (e-net, display, SCSI, etc) and it even uses PC type memory SIMMs to keep the cost down.
krowitz@RICHTER.MIT.EDU (David Krowitz) (06/01/91)
Yup. I've run a thin-net with Apollo token-ring cables for a limitted time over a limited distance (one room). It seemed to work. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)
ced@apollo.hp.com (Carl Davidson) (06/06/91)
Good morning (at least it's morning here), I've been following the thread summarized by David Funk's recent posting (enclosed in part below) and have permission from HP Marketing to respond, so here goes. I've excerpted David's posting and deleted some text but I haven't modified any of David's text. From article <9105310352.AA03763@icaen.uiowa.edu>, by dbfunk@ICAEN.UIOWA.EDU (David B Funk): > > Ok folks, rumor control time. ATR is not an option on the 700 series > WHEN RUNNING HP-UX. HP-UX has no idea how to talk to any kind of token > ring board. > However ATR -IS- an option for the 700 series under OSF. HP/Apollo > has said time & again that they -are- going to support OSF on ATR; > even to the extent of running both Domain/OS & OSF machines on the > same net. The 700 series hardware can run an ATR, it is just a matter > of the OS software drivers. > Everything David says here is absolutely correct. No "ifs", "ands", or "buts". It is important to note, however, that supporting ATR is NOT, I repeat NOT, equivalent to supporting the Domain file system OR the full suite of Domain network management tools on OSF nodes. This we are NOT doing. *********************************************************************** *********************************************************************** THE ONLY PROTOCOL WE WILL SUPPORT OVER ATR is IP. *********************************************************************** *********************************************************************** HP/OSF1 will treat ATR as "just another link". It will offer the same set of services over ATR that it offers over the built-in Ethernet controller (which, by the way, is an Intel 82596). If you have both the Ethernet and the ATR interfaces up, your HP/OSF1 node will route IP packets between the two networks but it will NOT route DDS packets. In order to facilitate interoperation with Domain nodes on ATR (why else would you want ATR on a Snake, right?), we are using the Domain Distributed Services (DDS) packet header as our frame header. This allows me to use the "r" commands, telnet, ftp, ping, and NFS between the Snake on my desk and any Domain/OS node on the same ATR or accessible through the IP Internet without any change to the Domain nodes' software. Of course, Snakes on ATR can also talk to other Snakes on ATR. In order to maintain the integrity of the ring, HP/OSF1 ATR will respond to "lcnode" (asknode_$who) packets and regenerate the request in order to allow the Domain/OS node performing the lcnode to wee the entire ring. If HP/OSF1 didn't do this, the lcnode output would stop at the node immediately "upstream" from the first Snake in the ring. We will also respond to the "bldt" command to the extent that a "bldt Net.Node_ID" done to an HP/OSF1 node from a Domain/OS node will return the string elicited by the OSF command "uname -a". We are doing this to allow people to figure out why a node shows up on an "lcnode", but can't be catalogued or otherwise accessed with Domain services. > > All the 700 series machines can take at least 1 EISA bus card. > The 750 has a cage with 4 EISA slots, the 730 has 1 standard EISA slot, > and the 720 has one optional EISA slot. > Again, all true. > > On the 720 it is down by > the power supply and if you don't order the optional adapter for it, > then it is hidden by a cover plate. > I think what David means is that if you don't order the optional adapter, then the space it would have taken up is covered by a blank plate. If so, then he is correct. > > Given that HP/Apollo already has a source of ISA bus ATR boards > I doubt that they need to go to somebody else for them. Maybe they are > looking for a new source of IBM token ring boards (16 megabit maybe?). > The board we are using is the latest revision of the board that is used with the DPCI/Ring product. It will get the node ID from a prom on the ATR board (if you have a 750, you can have two boards, in which case you'll have two proms) and will not have a boot prom installed (what good would it do). > > We are ordering ATR with our 750 ;). Our salesman did not have the > part number for it in his current catalog but he assures us he will find it, > even if he has to call Chelmsford. (no we are not buying it to run HP-UX ;). > > Thanks for your vote of confidence. I hope this clears up any remaining questions about ATR support on HP/OSF1 and Snakes. If there are any lingering questions, it would probably be best to send me e-mail rather than post them. Sometimes I don't have an opportunity to read news for a few days and your message could expire before I see it. I'll post responses to mail I receive to comp.sys.apollo as appropriate. I suppose this would be the place to say that even though I have permission from HP Marketing to post this I am not an official spokesman for HP and therefore if what I have written conflicts with some official policy or announcement either now or in the future, then, of course, the official policy or announcement is the one to believe. On the other hand, you can have a fairly high degree of confidence that the above is correct, since I'm the one writing the code and I currently have it running on my Snake here in Chelmsford. Best regards, -- Carl Davidson (508) 256-6600 x4361 | "What is the Existential The Apollo Systems Divison of | Vaccuum and does it come The Hewlett-Packard Company | with attachments?" DOMAIN: ced@apollo.HP.COM |
thompson@PAN.SSEC.HONEYWELL.COM (John Thompson) (06/07/91)
Carl -- > > Ok folks, rumor control time. ATR is not an option on the 700 series > > ... > > However ATR -IS- an option for the 700 series under OSF. HP/Apollo > > Everything David says here is absolutely correct. No "ifs", "ands", > or "buts".... > THE ONLY PROTOCOL WE WILL SUPPORT OVER ATR is IP. I hope that the rumours didn't start w/ the USENET letter. I tried to be clear in the fact that AT FIRST we wouldn't be able to have the Snakes in the same network. If we wait until Q?, we can of course run OSF, and that is of course true (unless HP decides to change their mind again). > In order to facilitate interoperation with Domain nodes on ATR (why else > would you want ATR on a Snake, right?).... Oh, I don't know. Maybe the large investment in building wiring; maybe the fact that ATR is faster than ethernet; maybe the fact that token-ring doesn't break down at high load levels; maybe because ATR is superior! > In order to maintain the integrity of the ring, HP/OSF1 ATR will respond to > "lcnode" (asknode_$who) packets and regenerate the request in order to > allow the Domain/OS node performing the lcnode to wee the entire ring. > If HP/OSF1 didn't do this, the lcnode output would stop at the node > immediately "upstream" from the first Snake in the ring. Good. Thank you. > We will also respond to the "bldt" command to the extent that a > "bldt Net.Node_ID" done to an HP/OSF1 node from a Domain/OS node will > return the string elicited by the OSF command "uname -a". We are doing this > to allow people to figure out why a node shows up on an "lcnode", but > can't be catalogued or otherwise accessed with Domain services. Better. Thank you even more. > I suppose this would be the place to say that even though I have permission > from HP Marketing to post this I am not an official spokesman for HP and > therefore if what I have written conflicts with some official policy or > announcement either now or in the future, then, of course, the official > policy or announcement is the one to believe. On the other hand, you can > have a fairly high degree of confidence that the above is correct, since > I'm the one writing the code and I currently have it running on my Snake > here in Chelmsford. I realize that it's necessary, but I hope that, after you re-read the above paragraph, you understand some of our concerns. HP has promised longer support for Domain/OS and yanked it; HP has promised a PRISM-II and yanked it (I know many of the details, and can't (quite) blame mgmt for the problem); HP has promised Domain/OS on the first merged RISC box and yanked it; HP has promised OSF support on all DN3000+ boxes and yanked it. I know that you are writing only what is true and correct, but HP management seems to revel in changing the truth to suit their whims. IBIWISI!!! (I'llBelieveItWhenISeeIt). Regards -- jt -- John Thompson Honeywell, SSEC Plymouth, MN 55441 thompson@pan.ssec.honeywell.com When in danger, when in doubt -- run in circles, scream and shout.