1 INFO-VAX	Wed, 21 Jun 2006	Volume 2006 : Issue 342       Contents:) console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch - Re: console message involving RUN/STOP switch  Hung System RWMBX state  RE: Hung System RWMBX state  Re: Hung System RWMBX state  RE: Hung System RWMBX state  Re: Hung System RWMBX state  Re: Hung System RWMBX state  Re: Hung System RWMBX state  Re: Hung System RWMBX state 1 OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... 5 Re: OT: IA64's speed beaten by orders of magnitude... + Re: Power supply for Alphastation 400 4/233 + Re: Power supply for Alphastation 400 4/233 * Setting up my NIC for Gb Ethernet traffic.. Re: Setting up my NIC for Gb Ethernet traffic.  F ----------------------------------------------------------------------  + Date: Tue, 20 Jun 2006 19:41:51 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)2 Subject: console message involving RUN/STOP switch$ Message-ID: <e79j1v$tqe$1@online.de>  E I'm testing out some old hardware, an ALPHAserver 2100.  It powers up - fine, the console sees all the devices etc.     6 I wanted to boot it off of a CD, but I get the message  K    dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch     failed to open dka600.6.0.1  ? I tried a 7.3-2 CD and a 7.3-1 CD.  Both are copies, but I have F successfully booted off of them in other machines in the past.  The CDE drive in the 2100 is an RRD43.  I have an original 7.1 CD but I can't  find it at the moment.  H If all else fails, I'll try to boot it off of a system disk from anotherG machine.  (However, since this is configured for a cluster etc, I can't = expect to boot straight off of it; I'll probably have to do a  conversational boot.)   E Right now, I just want to see the machine come up with VMS to verify   that everything is working OK.  F Could it be that I get the error message since I tried copies and not H original CDs?  (I HAVE booted off of these on other ALPHAs in the past,  with IIRC also an RRD43.)   I What is the RUN/STOP switch?  Is it some physical switch on the machine,  ( or some console command I have to enter?   ------------------------------  + Date: Tue, 20 Jun 2006 19:47:24 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: console message involving RUN/STOP switch$ Message-ID: <e79jcc$tqe$2@online.de>  D In article <e79j1v$tqe$1@online.de>, helbig@astro.multiCLOTHESvax.de3 (Phillip Helbig---remove CLOTHES to reply) writes:    E Another 2100 question: There are 8 SBB slots for disk drives.  At the F console, I see devices DRA0, DRA1,...DRA7.  Do these correspond to theG disk slots?  If so, what kind of disk can I put in (hopefully something 0 like an RZ28.)  If not, what are the DR devices?   ------------------------------  % Date: Tue, 20 Jun 2006 22:24:14 +0200 # From: Kjell Andresen <kjell@dod.no> 6 Subject: Re: console message involving RUN/STOP switch+ Message-ID: <u8n64ivzic1.fsf@modell.uio.no>   R helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes:  G > I'm testing out some old hardware, an ALPHAserver 2100.  It powers up / > fine, the console sees all the devices etc.    > 8 > I wanted to boot it off of a CD, but I get the message > M >    dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch   >    failed to open dka600.6.0.1  = There are 3 buttons on the 2100-panel; power, halt and reset. = If the halt button is pressed you need to press it once more   to get the box booted.  K > What is the RUN/STOP switch?  Is it some physical switch on the machine,  * > or some console command I have to enter?   Yes, it's a physical button.@ If I remember correct it is a circle with a triangle upside down inside.    Regards,A Kjell Andresen  Systems administrator, University of Oslo, Norway >                 Center for Information Technology Services and)                 Department of Geosciences -                 email: drift@geofysikk.uio.no    ------------------------------  # Date: Tue, 20 Jun 2006 20:50:02 GMT ( From: "P. Thompson" <no-spam@new.rr.com>6 Subject: Re: console message involving RUN/STOP switchD Message-ID: <Pine.LNX.4.64.0606201547570.7036@localhost.localdomain>  * On Tue, 20 Jun 2006, Kjell Andresen wrote:  N >>    dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch! >>    failed to open dka600.6.0.1  >  > Yes, it's a physical button.B > If I remember correct it is a circle with a triangle upside down	 > inside.  > G I think he's confusing halt with the run/stop which can be cabled into  ( most SCSI disks (and typically are not).D My guess is that your copies of media are written on media that the J ancient drive on the 2100 can't recognize.  I've run into that on various D systems where CDRW writes just differently enough on the media that # ancient relics can't quite read it. I Best bet is to cable in a newer 512 byte sector scsi CD drive and see if  	 it works.   I On an unrelated note DRA drives are hooked to an onboard RAID controller.    ------------------------------  + Date: Tue, 20 Jun 2006 21:39:27 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: console message involving RUN/STOP switch$ Message-ID: <e79puf$bko$2@online.de>  H In article <Pine.LNX.4.64.0606201547570.7036@localhost.localdomain>, "P.' Thompson" <no-spam@new.rr.com> writes:    F > My guess is that your copies of media are written on media that the . > ancient drive on the 2100 can't recognize.     Yes, that seems to be the case.   K > Best bet is to cable in a newer 512 byte sector scsi CD drive and see if   > it works.   A I get a bit further there (see my other response in this thread).   K > On an unrelated note DRA drives are hooked to an onboard RAID controller.   F That's what I suspected.  So one could just stick in any SBB SCSI disk& of appropriate vintage (say an RZ28)?    ------------------------------  + Date: Tue, 20 Jun 2006 21:42:35 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)6 Subject: Re: console message involving RUN/STOP switch$ Message-ID: <e79q4b$bko$3@online.de>  C helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to  reply) writes:    G > I'm testing out some old hardware, an ALPHAserver 2100.  It powers up / > fine, the console sees all the devices etc.    > 8 > I wanted to boot it off of a CD, but I get the message > M >    dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch   >    failed to open dka600.6.0.1 > M >    dka600.6.0.1 has no media present or is disabled via the RUN/STOP switch   >    failed to open dka600.6.0.1 >   A > I tried a 7.3-2 CD and a 7.3-1 CD.  Both are copies, but I have H > successfully booted off of them in other machines in the past.  The CDG > drive in the 2100 is an RRD43.  I have an original 7.1 CD but I can't  > find it at the moment.  I I'm pretty sure now that it is due to the fact that I am using copies of  H CDs and not original CDs.  (Unfortunately, I can't find my original 7.1 I CD.)  In terms of data, my copies are exact copies.  The problem is that  B the original CDs have small pits for the bits and the copies have F different reflectivity.  Old drives can't read copies.  Some somewhat E newer drives can read them but they won't work as a boot drive.  New  G drives don't have any problems at all.  (I have installed and upgraded  # VMS many times from copies of CDs.)   G I have an expansion box with 5 TOSHIBA CD drives in it.  Since the 2100 G has two tape drives and a CD drive, that is one device too many.  I set @ the IDs of the CDs in my box such that there is only one SCSI-IDH conflict.  When I do this, and put the CD in one of the drives in my boxF (connected to the SCSI connector near the top of the 2100, not the oneG on the card in the expansion slot---I don't see anything at the console D if I connect to it), the machine starts to boot, but stops after the message       jumping to bootstrap code  H Like I said, there are two devices with the same SCSI ID on the bus and @ also the cable is quite long.  Either or both of these could be H contributing to the problem (perhaps along with the fact that the drive I is perhaps not completely able to read the copy CD, though I'm sure that  2 it can handle it if it is not being used to boot).  G Unfortunately, trying a shorter cable would mean getting the heavy box  D down off of a high shelf and removing one of the devices on the bus E would mean opening up the expansion box or the 2100 itself.  I don't   have time to test this now.   E Any serious hardware errors would have showed up at the console or on E the alphanumeric display on the machine itself.  The fact that it got @ this far indicates to me that there is no real problem with it.    ------------------------------  % Date: Tue, 20 Jun 2006 14:30:19 -0400 1 From: "Farrell, Michael" <MFarrell@Voltdelta.com>   Subject: Hung System RWMBX stateL Message-ID: <085BCCCF596B684092B66310B1D3BA7D02CC6932@NJ103EX1.EAST.VIS.COM>   Help!   H I have one node of my Alpha Cluster with more than half of its processes in RWMBX state.   > We have no active sessions on that system so we cannot get in.  H The QUEUE MANAGER is running on that system and process QUEMAN is one ofG those in RWMBX state.  So we can't determine the state of the 273 batch G jobs running on that system and the 13 running on the ES47 that is also  in the cluster.   F Is there any way of saving this system without crashing and rebooting?  B Its an ES45 running VMS 7.3-2.  All systems in the cluster are VMS 7.3-2.  D To my knowledge the system is patched with all patches up to but not# included those now being announced.   0 Does anyone have any idea if this can be helped?   TIA    Mike Farrell   ------------------------------  % Date: Tue, 20 Jun 2006 14:34:20 -0400 1 From: "Farrell, Michael" <MFarrell@Voltdelta.com> $ Subject: RE: Hung System RWMBX stateL Message-ID: <085BCCCF596B684092B66310B1D3BA7D02CC6940@NJ103EX1.EAST.VIS.COM>   A P.S. to this message:   F There is also a process running on this system named VPM_14100026 thatF is in COM and has used far more CPU time than any other process on the system.   G I don't know what that process is.  There is not process with that or a D similar name running on any of my other systems.  Don't know if that! factors into this problem or not.    TIA    Mike Farrell   -----Original Message----- From: Farrell, Michael=20 $ Sent: Tuesday, June 20, 2006 2:30 PM To: 'Info-VAX@Mvb.Saic.Com'  Cc: 'HP-G.Peleg'  Subject: Hung System RWMBX state   Help!   H I have one node of my Alpha Cluster with more than half of its processes in RWMBX state.   > We have no active sessions on that system so we cannot get in.  H The QUEUE MANAGER is running on that system and process QUEMAN is one ofG those in RWMBX state.  So we can't determine the state of the 273 batch G jobs running on that system and the 13 running on the ES47 that is also  in the cluster.   F Is there any way of saving this system without crashing and rebooting?  B Its an ES45 running VMS 7.3-2.  All systems in the cluster are VMS 7.3-2.  D To my knowledge the system is patched with all patches up to but not# included those now being announced.   0 Does anyone have any idea if this can be helped?   TIA    Mike Farrell   ------------------------------    Date: 20 Jun 2006 12:57:52 -0700" From: dave.baxter@bannerhealth.com$ Subject: Re: Hung System RWMBX stateC Message-ID: <1150833472.639124.323050@h76g2000cwa.googlegroups.com>    Mike, @        Can you access the system from another cluster node usingE sysman???   If you can, you might be able to do something from there. D   The "VPM" process you mention is probably something to do with theF monitor utility (i.e. VPM_SERVER sends monitor over TCPIP (I believe))% and it can probably be safely killed.    Dave     Farrell, Michael wrote:  > Help!  > J > I have one node of my Alpha Cluster with more than half of its processes > in RWMBX state.  > @ > We have no active sessions on that system so we cannot get in. > J > The QUEUE MANAGER is running on that system and process QUEMAN is one ofI > those in RWMBX state.  So we can't determine the state of the 273 batch I > jobs running on that system and the 13 running on the ES47 that is also  > in the cluster.  > H > Is there any way of saving this system without crashing and rebooting? > D > Its an ES45 running VMS 7.3-2.  All systems in the cluster are VMS > 7.3-2. > F > To my knowledge the system is patched with all patches up to but not% > included those now being announced.  > 2 > Does anyone have any idea if this can be helped? >  > TIA  >  > Mike Farrell   ------------------------------  % Date: Tue, 20 Jun 2006 16:25:08 -0400 1 From: "Farrell, Michael" <MFarrell@Voltdelta.com> $ Subject: RE: Hung System RWMBX stateL Message-ID: <085BCCCF596B684092B66310B1D3BA7D02CC6ABE@NJ103EX1.EAST.VIS.COM>   Dave,   	 Thanks=20   B We could not do that because there were no mailboxes available forF logging in.  I lost the only session I had when I did an "anal/sys" onC the QUEMAN process.  According to Mark Jilson, who helped me, I was  locked out at that point.   G We ended up rebooting and took a crash dump for further analysis to see  how this happened.   Mike  B P.S.	You're right.  The "VPM" process was a monitor process but we% could not kill it.  It would not die.    -----Original Message-----H From: dave.baxter@bannerhealth.com [mailto:dave.baxter@bannerhealth.com]  $ Sent: Tuesday, June 20, 2006 3:58 PM To: Info-VAX@Mvb.Saic.Com $ Subject: Re: Hung System RWMBX state   Mike, @        Can you access the system from another cluster node usingE sysman???   If you can, you might be able to do something from there. D   The "VPM" process you mention is probably something to do with theF monitor utility (i.e. VPM_SERVER sends monitor over TCPIP (I believe))% and it can probably be safely killed.    Dave     Farrell, Michael wrote:  > Help!  > @ > I have one node of my Alpha Cluster with more than half of its	 processes  > in RWMBX state.  > @ > We have no active sessions on that system so we cannot get in. > G > The QUEUE MANAGER is running on that system and process QUEMAN is one  ofC > those in RWMBX state.  So we can't determine the state of the 273  batch D > jobs running on that system and the 13 running on the ES47 that is also > in the cluster.  > H > Is there any way of saving this system without crashing and rebooting? > D > Its an ES45 running VMS 7.3-2.  All systems in the cluster are VMS > 7.3-2. > F > To my knowledge the system is patched with all patches up to but not% > included those now being announced.  > 2 > Does anyone have any idea if this can be helped? >=20 > TIA  >=20 > Mike Farrell   ------------------------------    Date: 20 Jun 2006 13:24:58 -0700) From: "Ken Robinson" <kenrbnsn@gmail.com> $ Subject: Re: Hung System RWMBX stateB Message-ID: <1150835098.178958.41970@b68g2000cwa.googlegroups.com>  # dave.baxter@bannerhealth.com wrote:  > Mike, B >        Can you access the system from another cluster node usingG > sysman???   If you can, you might be able to do something from there. F >   The "VPM" process you mention is probably something to do with theH > monitor utility (i.e. VPM_SERVER sends monitor over TCPIP (I believe))' > and it can probably be safely killed.   A You don't need to use SYSMAN for this. From a suitably privileged A account one another node in the cluster you can either reduce the ? priority of the run away process or just kill it, using the set $ process/id and the stop/id commands.  C If you use a console manger to monitor your consoles, then that may  still be loged in.   Ken    ------------------------------  % Date: Tue, 20 Jun 2006 16:28:07 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Re: Hung System RWMBX state, Message-ID: <44985A56.500C0280@teksavvy.com>   "Farrell, Michael" wrote: J > I have one node of my Alpha Cluster with more than half of its processes > in RWMBX state.     B RWMBX means that a process is (generally) attempting to write to aC mailbox that is not being read, so the mailbox buffer gets full and 4 anyone trying to write to it then goes in the RWMBX.  F Do all the processes in that state have anything in common ? (runs the  same application for instance ?)  ' Pick one process that is in that state.    $ANA/SYSTEM  $SET PROCESS/ID=xxxxxx $SHOW PROC/CHANNELS   5 You should see one or more mailbox devices. (MBAnnn:)   B What you can do to clear one process is: (in a spawned subprocess)   $OPEN/READ temp MBAnnn:  $READ temp buffer   * repeat the read until it doesn't complete.  H If there are multiple mailboxes used by the process, some may jam at theG first read (no data to read). You can then kill the subprocess and then  try the next mailbox.   H Consider a case where a server process reads data from a client and then writes data back to client.   F If, in writing data to the server, the client enters RWMBX, the clientE will not be able to complete any reads from the response mailbox, and C the server will then enter RWMBX and you've got yourself an italian  traffic jam.  C By manually emptying the mailbox devices, you could get the traffic G moving again, but the clients will not be "right" because some may then H wait for a responmse from teh server which will never come. But at least? you shoudl then be able to STOP/ID those processes that are not " responsive and not have to reboot.    E RWMBX is one of the easiest/cleanest of the nasty RW disasters on VMS E because you can manipulate the mailboxes from DCL and remove any data  that is blocking the pipes.    ------------------------------  % Date: Tue, 20 Jun 2006 17:03:51 -0400 . From: JF Mezei <jfmezei.spamnot@vaxination.ca>$ Subject: Re: Hung System RWMBX state- Message-ID: <449862B2.39192ADA@vaxination.ca>    "Farrell, Michael" wrote: D > We could not do that because there were no mailboxes available for > logging in.   F Can you elaborate on exactly what would happen if you tried to login ?$ How far into the login would it go ?  G Did you try /NOCOMMAND in the login to bypass the sylogin and login.com A ? (if those are where some mailbox was involved due to some local * commands, you could have still logged in).   ------------------------------  % Date: Tue, 20 Jun 2006 17:21:21 -0700 ( From: Jeff Cameron <roktsci@comcast.net>$ Subject: Re: Hung System RWMBX state0 Message-ID: <C0BDDF11.21AEB%roktsci@comcast.net>  K I cant remember which memory resource is used for buffered mailboxes, and I I don't have access to my VMS system just now, but there is a resource that 6 might help fix the mailbox problem. Have you looked at $SHO MEM/POOL/ALL/FULL* To see if you are running low on anything?J Can anyone else recall which memory resource I'm thinking of? Is it PPAGED DYNAMIC MEMORY?    Certainly AUTOGEN might help.    Jeff   On 6/20/06 11:30 AM, in article I 085BCCCF596B684092B66310B1D3BA7D02CC6932@NJ103EX1.EAST.VIS.COM, "Farrell, ( Michael" <MFarrell@Voltdelta.com> wrote:   > Help!  > J > I have one node of my Alpha Cluster with more than half of its processes > in RWMBX state.  > @ > We have no active sessions on that system so we cannot get in. > J > The QUEUE MANAGER is running on that system and process QUEMAN is one ofI > those in RWMBX state.  So we can't determine the state of the 273 batch I > jobs running on that system and the 13 running on the ES47 that is also  > in the cluster.  > H > Is there any way of saving this system without crashing and rebooting? > D > Its an ES45 running VMS 7.3-2.  All systems in the cluster are VMS > 7.3-2. > F > To my knowledge the system is patched with all patches up to but not% > included those now being announced.  > 2 > Does anyone have any idea if this can be helped? >  > TIA  >  > Mike Farrell >    ------------------------------  % Date: Tue, 20 Jun 2006 17:56:07 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: OT: IA64's speed beaten by orders of magnitude..., Message-ID: <44986EEF.970216A9@teksavvy.com>  a http://news.com.com/Chip+breaks+speed+record+in+deep+freeze/2100-1006_3-6085568.html?tag=nefd.top     C IBM and Georgia tech have built a chip that runs at 350 Ghz. And by F dipping in in liquid helium at near 0 kelvin, they were able to boost its speed to nearly 500 Ghz.  < Makes my all mighty MV II look rather pale in comparison :-)  G So, when will Intel's IA64 thing reach 350 Ghz speeds ? :-) ;-) :-) ;-)  :-) :-) :-) :-)     @ Out of curiosity, since they use crystals to generate the timingE signals, is there a phsyical limit on how fast a clock can be made to 
 oscilate ?   ------------------------------  % Date: Tue, 20 Jun 2006 18:38:51 -0700 # From: "Tom Linden" <tom@kednos.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude...) Message-ID: <op.tbg2u1vjzgicya@hyrrokkin>   . On Tue, 20 Jun 2006 14:56:07 -0700, JF Mezei  % <jfmezei.spamnot@teksavvy.com> wrote:   c > http://news.com.com/Chip+breaks+speed+record+in+deep+freeze/2100-1006_3-6085568.html?tag=nefd.top  >  > E > IBM and Georgia tech have built a chip that runs at 350 Ghz. And by H > dipping in in liquid helium at near 0 kelvin, they were able to boost > its speed to nearly 500 Ghz. > > > Makes my all mighty MV II look rather pale in comparison :-) > I > So, when will Intel's IA64 thing reach 350 Ghz speeds ? :-) ;-) :-) ;-)  > :-) :-) :-) :-)  > 3 When hell freezes over, which BTW, happens at -451K  > B > Out of curiosity, since they use crystals to generate the timingG > signals, is there a phsyical limit on how fast a clock can be made to  > oscilate ?   ------------------------------  % Date: Tue, 20 Jun 2006 22:39:58 -0400 ' From: Dave Froble <davef@tsoft-inc.com> > Subject: Re: OT: IA64's speed beaten by orders of magnitude...9 Message-ID: <qrOdnehQ6PHsLQXZnZ2dnUVZ_vednZ2d@libcom.com>    Tom Linden wrote: / > On Tue, 20 Jun 2006 14:56:07 -0700, JF Mezei  ' > <jfmezei.spamnot@teksavvy.com> wrote:  > e >> http://news.com.com/Chip+breaks+speed+record+in+deep+freeze/2100-1006_3-6085568.html?tag=nefd.top   >> >> >>F >> IBM and Georgia tech have built a chip that runs at 350 Ghz. And byI >> dipping in in liquid helium at near 0 kelvin, they were able to boost  >> its speed to nearly 500 Ghz.  >>? >> Makes my all mighty MV II look rather pale in comparison :-)  >>J >> So, when will Intel's IA64 thing reach 350 Ghz speeds ? :-) ;-) :-) ;-) >> :-) :-) :-) :-) >>5 > When hell freezes over, which BTW, happens at -451K   I If my memory is working at all, I seem to remember that the Kelvin scale  B is absolute.  The temperature you're thinking about would be zero ? degrees Kelvin.  What's called 'absolute zero', no heat at all.   & Nice comment about the itanic, though.  C >> Out of curiosity, since they use crystals to generate the timing H >> signals, is there a phsyical limit on how fast a clock can be made to
 >> oscilate ?  >      --  4 David Froble                       Tel: 724-529-0450> Dave Froble Enterprises, Inc.      E-Mail: davef@tsoft-inc.com DFE Ultralights, Inc.  170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 20 Jun 2006 16:35:55 -0400 / From: "William Webb" <william.w.webb@gmail.com> 4 Subject: Re: Power supply for Alphastation 400 4/233I Message-ID: <8660a3a10606201335j104dc37chc1a3d9efe3838d3c@mail.gmail.com>   - On 6/19/06, Z <SpamDumpst3r@yahoo.com> wrote: M > Is the power supply in an Alphastation 400 4/233 a vanilla AT power supply?  >   D None of the AlphaStations I can think of use a COTS AT power supply.  D The connectors are different, an idea either copied by or from Dell.   Sorry    WWWebb --  C NOTE: This email address is only used for noncommerical VMS-related  correspondence. C All unsolicited commercial email will be deemed to be a request for 8 services pursuant to the terms and conditions located at# http://bellsouthpwp.net/w/e/webbww/    ------------------------------  % Date: Tue, 20 Jun 2006 21:51:53 -0500 6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler>4 Subject: Re: Power supply for Alphastation 400 4/233@ Message-ID: <craigberry-EFD72F.21515320062006@free.teranews.com>  E In article <JWElg.1956$z75.250@fe04.lga>, Z <SpamDumpst3r@yahoo.com>   wrote:  M > Is the power supply in an Alphastation 400 4/233 a vanilla AT power supply?   H No.  It has more connector thingies.  Forgive my imprecise response but ) that's what I can remember at the moment.    --  = Posted via a free Usenet account from http://www.teranews.com    ------------------------------    Date: 20 Jun 2006 13:46:22 -0700" From: dave.baxter@bannerhealth.com3 Subject: Setting up my NIC for Gb Ethernet traffic. C Message-ID: <1150836382.941005.155940@g10g2000cwb.googlegroups.com>   @       I am trying to acheive Gb-level throughput from TSM clientG (ABClient) running on my VMS system (ES40) to a MainFrame (OS390) which F is the TSM Server.      My VMS Version is 7.3-2, and my TCPIP stack isE TCPIP Services for OpenVMS Version 5.4, ECO 5, (+ the 5+8 patch kit).   A      My NIC is a 10/100/1000 Adapter and I am linking up with the @ network switch OK at that speed and with full duplex enabled and> functioning.      I have enable JUMBO Frames on this NIC using   $ mc lancp set dev /jumbo EWB0  F I do this right after DECNet starts, but before TCPIP Starts,  Is this the correct place to do this.    $ mc lancp show dev ewb0 /char    Device Characteristics EWB0: '                   Value  Characteristic '                   -----  -------------- +                    1500  Device buffer size (                  Normal  Controller mode/                External  Internal loopback mode -       00-10-18-10-00-5E  Hardware LAN address /                          Multicast address list -                 CSMA/CD  Communication medium ,       FF-FF-FF-FF-FF-FF  Current LAN address0                     128  Minimum receive buffers0                     256  Maximum receive buffers*                    Yes  Full duplex enable0                     Yes  Full duplex operational(             TwistedPair  Line media type*                    1000  Line speed (mbps)*     Disabled/No Failset  Logical LAN state*                       0  Failover priority LANCP>  B I am concerned that the Device buffer size shown above still reads= 1500.      Is this to be expected for what I am trying to do.   E Viewing the interface with ifconfig seems to indicate that we are set  up OK, "ipmtu 9000"     MILL03$SYSMGR>> ifconfig we1WE1:F flags=c43<UP,BROADCAST,RUNNING,MULTICAST,SIMPLEX>    *inet xx.xx.x.xxx3 netmask ffff0000 broadcast xx.xx.255.255 ipmtu 9000     G      Basically, I am trying to confirm that I have done everything that ? I need to do on my own system, before I start pointing fingers. D My network people tell me that they have enabled Jumbo Frames on all1 the devices in the path from me to the mainframe.   1 Does anyone see anything that I might be missing.   E e.g. I noticed that there are some Console Environment variables that > appear to relate to NIC's, can anyone shed any light on those.   thanks     Dave.    ------------------------------  # Date: Tue, 20 Jun 2006 21:50:09 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> 7 Subject: Re: Setting up my NIC for Gb Ethernet traffic. = Message-ID: <l4_lg.89336$wl.39265@text.news.blueyonder.co.uk>   J I suspect that you need to enable Jumbo frames before starting DECnet (and0 maybe before starting SCS if this is a cluster).  H The SYSGEN parameter LAN_FLAGS (bit 6) should enable Jumbo frames. LANCPJ DEFINE might be a good thing too. Unfortunately my test systems are out of- service at the moment, so I can't check this.    --     Hope this helps, Colin. ) colin DOT butcher AT xdelta DOT co DOT uk E It's not mine, but I like this definition: Legacy = stuff that works.    ------------------------------   End of INFO-VAX 2006.342 ************************