1 INFO-VAX	Mon, 29 Oct 2001	Volume 2001 : Issue 601       Contents: Re: EMC Symmetrix on OpenVMS ? Re: EMC Symmetrix on OpenVMS ?! Re: Single or Multiple Sys Disks?  Re: SSH for Alpha VMS # Re: Windows XP reality check please # Re: Windows XP reality check please   F ----------------------------------------------------------------------  % Date: Sun, 28 Oct 2001 18:39:27 -0500 * From: Chuck Chopp <ChuckChopp@rtfmcsi.com>' Subject: Re: EMC Symmetrix on OpenVMS ? + Message-ID: <3BDC972F.3BBC21BB@rtfmcsi.com>    Carlos Chua wrote:   > Hi,  > M > Has anyone tried connecting their Alphaserver to EMC Symmetrix? My OS level N > is OpenVMS V6.2. As this doesn't support multiple path access to disks driveH > yet, I need some suggestion on how to create a configuration that willK > prevent single point of failure. Is there anyway to protect from the SCSI F > HBA failure aside from using volume shadowing? EMC disks are alreadyL > mirrored protected. If I use volume shadowing, I will end up with 4 copies > which is too much. >  > Thanks in advance, >  > Carlos > chuacarlos@hotmail.com  O Arghhh!  I really dislike it when memory no longer serves as it should...  I've I spent way too much time the past few months on NetWare and Win32 platform ? projects.  I've been missing out on some quality time with VMS.    Anyway, here goes...  G If you use CMD Trident controllers as the front-end you can effectively K eliminate the single point of failure.  However, IIRC, you'll find that the L Trident controllers only hook up to a CI bus and thus you'll lose out on theP performance that you would otherwise get from attaching the AlphaServer directlyP to the EMC Symmetrix server via a SCSI3 controller.  One of my clients does thisP right now and it allows them to access storage capacity on the EMC server from aM mixed architecture cluster consisting of 2 VAX systems [6000 and 7000 series] M and of 2AlphaServer systems [4000 and 7000 series].  It works OK for them w/o I any problems, but their I/O throughput needs do not exceed what the CI is  capable of delivering to them.  L Also, I vaguely remember that somewhere in either v6.x or v7.0 of OpenVMS anP option appeared to allow multiple SCSI paths to the same disk and that you couldM forcibly rename a SCSI device so that it had the same alloclass, controller & O unit # regardless of which SCSI controller it was being accessed through.  This N would let you connect the same host to the same EMC server via 2 separate SCSIN controllers and still have the disk look like a single device.  Perhaps one ofL the folks who has actually used this feature could fill in some more detailsH about how it is configured and what versions of OpenVMS support its use.     Regards,   Chuck  -- Chuck Chopp   8 ChuckChopp@rtfmcsi.com            http://www.rtfmcsi.com0                                   ICQ # 22321532@ RTFM Consulting Services Inc.     864 801 2795 voice & voicemail2 103 Autumn Hill Road              864 801 2774 fax4 Greer, SC  29651                  800 774 0718 pager7                                   8007740718@skytel.com    ------------------------------  # Date: Mon, 29 Oct 2001 05:54:00 GMT * From: Lee Y T Mah <lytmah@telusplanet.net>' Subject: Re: EMC Symmetrix on OpenVMS ? / Message-ID: <3BDCEEF7.A75DC087@telusplanet.net>   T A word of warning.  Using third party storage on VMS systems may negate some aspectsT of your Compaq support (E.g., HBSV).  Also, Cerner, a large player in the VMS healthT care field, may not provide support for problems resulting from third-party storage.   Chuck Chopp wrote:   > Carlos Chua wrote: >  > > Hi,  > > O > > Has anyone tried connecting their Alphaserver to EMC Symmetrix? My OS level P > > is OpenVMS V6.2. As this doesn't support multiple path access to disks driveJ > > yet, I need some suggestion on how to create a configuration that willM > > prevent single point of failure. Is there anyway to protect from the SCSI H > > HBA failure aside from using volume shadowing? EMC disks are alreadyN > > mirrored protected. If I use volume shadowing, I will end up with 4 copies > > which is too much. > >  > > Thanks in advance, > > 
 > > Carlos > > chuacarlos@hotmail.com > Q > Arghhh!  I really dislike it when memory no longer serves as it should...  I've K > spent way too much time the past few months on NetWare and Win32 platform A > projects.  I've been missing out on some quality time with VMS.  >  > Anyway, here goes... > I > If you use CMD Trident controllers as the front-end you can effectively M > eliminate the single point of failure.  However, IIRC, you'll find that the N > Trident controllers only hook up to a CI bus and thus you'll lose out on theR > performance that you would otherwise get from attaching the AlphaServer directlyR > to the EMC Symmetrix server via a SCSI3 controller.  One of my clients does thisR > right now and it allows them to access storage capacity on the EMC server from aO > mixed architecture cluster consisting of 2 VAX systems [6000 and 7000 series] O > and of 2AlphaServer systems [4000 and 7000 series].  It works OK for them w/o K > any problems, but their I/O throughput needs do not exceed what the CI is   > capable of delivering to them. > N > Also, I vaguely remember that somewhere in either v6.x or v7.0 of OpenVMS anR > option appeared to allow multiple SCSI paths to the same disk and that you couldO > forcibly rename a SCSI device so that it had the same alloclass, controller & Q > unit # regardless of which SCSI controller it was being accessed through.  This P > would let you connect the same host to the same EMC server via 2 separate SCSIP > controllers and still have the disk look like a single device.  Perhaps one ofN > the folks who has actually used this feature could fill in some more detailsJ > about how it is configured and what versions of OpenVMS support its use. > 
 > Regards, >  > Chuck  > --
 > Chuck Chopp  > : > ChuckChopp@rtfmcsi.com            http://www.rtfmcsi.com2 >                                   ICQ # 22321532B > RTFM Consulting Services Inc.     864 801 2795 voice & voicemail4 > 103 Autumn Hill Road              864 801 2774 fax6 > Greer, SC  29651                  800 774 0718 pager9 >                                   8007740718@skytel.com    -- Lee    lytmah@telusplanet.net   ------------------------------  % Date: Sun, 28 Oct 2001 22:06:32 +0100 & From: John McLean <mcleanj@dplanet.ch>* Subject: Re: Single or Multiple Sys Disks?* Message-ID: <3BDC7358.37B7031C@dplanet.ch>   Ed Wilts wrote:  > ; > In article <3BD58F71.1E80C13C@vmmc.org>, "Jack Trachtman" " > <Jack.Trachtman@vmmc.org> wrote: > K > > After all these years managing stand alone VMS systems, I'll finally be ? > > getting to manage two clustered systems: 2xES45 and 3xES45.  > > F > > My question: How to choose between single system disk and multipleH > > disks.  The first is easier to manage, but the second allows rolling, > > upgrades w/less down time for the users. > N > I'd vote for 2 system disks, plus a separate common disk that will hold your, > sysuaf, rightslist, vmsmail profiles, etc. > H > My thinking is that you can always install patches on one system disk,D > and even if all hell breaks loose, you can mount the busted set onJ > another system and fix the darn thing.  Rolling upgrades become simpler,C > and although you have to do all your patches twice,  the benefits I > far outweigh the gains.  The first time you need to have another system X > up and running to fix a broken disk will be the day you vow never to limit yourself to > a single system disk.  > G > Also to consider:  the size of your crash dump files, whether they're E > common or individual (5 individual dump files on 1 system disk will I > probably fill your disk depending on how much memory you have), and how  > busy the system disks are. >  >         .../Ed >  > --  > Ed Wilts, Mounds View, MN, USA > mailto:ewilts@ewilts.org  E I missed the start of this thread but I'd opt for one system disk but G make it shadowed.  Upgrade by dropping one member out of the shadow set F to keep it in reserve in case the upgrade goes to mush and you have to reinstate the old system.   F As for dump files, since about VMS7.2 (I think) your dump files can be  on disks other than system disk.     John McLean    ------------------------------  % Date: Sun, 28 Oct 2001 21:20:51 -0600 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: SSH for Alpha VMS' Message-ID: <3BDCCB13.59548EC7@fsi.net>    Bob Ceculski wrote:  > Z > jamese@beast.dtsw.army.mil wrote in message news:<01102616433509@beast.dtsw.army.mil>...M > > bob@instantwhip.com (Bob Ceculski) wrote on 26 Oct 2001 13:00:15 -0700 in 6 > > <d7791aa1.0110261200.772d3e91@posting.google.com>: > > 9 > > > "Amy Lewis" <amylewis@pacbell.net> wrote in message D > > > news:<iNYB7.3138$Ou1.1034800247@newssvr21.news.prodigy.com>...3 > > > > "L Arut" <larut@yahoo.com> wrote in message ? > > > > news:73f00c8f.0105282321.13854644@posting.google.com... C > > > > > Is there anyway of getting sn SSH server pre-compiled for G > > > > > an Alpha VMS system since I don't have access to DEC C having 8 > > > > > surrendered our VAX C license many years ago ? > > > >  > > > > Amy Lewis  > > > M > > > run tcpware ... it performs alot better than either multinet or ucx ...  > > / > > Do you have any statistics to back this up?  > > > > > Ed James                           ed.james@telecomsys.com9 > > TeleCommunications Systems, Inc.   voice 410-295-1919 ? > > 2024 West Street, Suite 300              800-810-0827 x1919 9 > > Annapolis, MD 21401-3556           fax   410-280-1094  > G > i certainly do ... we went thru several months of performance testing J > on our alpha web server using ucx, multinet and tcpware ... we ran threeE > different web servers, osu, apache and purveyor ... tcpware ran the J > crispest hands down ... the same thing happened with ucx from vms kernelG > based 4 to inix kernel based version 5 ... it got "slower"!  our vice J > president took part in the testing an the response times were definitiveD > same for purveyor ... it was written for vms (processed based) ...D > even folks at process software agreed with me when i told them theH > results and said that they had done some adjustments to narrow the gapA > but admit it still exists ... properly written vms kernel based D > software will always out run unix based, thats common sense and weE > proved it ... both products are almost identical in features on vms E > except i love using decnet phase iv over ip, something no other vms H > ip stack can do ... so why not use "the" premier ip stack for vms that0 > was "written" for vms by former vms engineers?  $ IMO, the management interface sucks.  C Multinet has its drawbacks, also, but sucks much less than TCPware.   @ Both could have taken a lesson from NCP, without making the same mistakes as UCX.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Sun, 28 Oct 2001 23:54:39 GMT " From: Art Rice <arice@myhouse.org>, Subject: Re: Windows XP reality check please: Message-ID: <3V0D7.1142$3T6.585900@paloalto-snr2.gtei.net>   Art Rice wrote:   ( > paddy.o'brien@zzz.tg.nsw.gov.au wrote: >  >> Paul Anderson wrote:  >>  E >>>At least they've finally admitted that those who work with Windoze J >>>computers have been wasting a lot of time dealing with poor OS quality. >>  I >> And still waste a lot of time dealing with the newer version rubbish.  : >> And the incomptabilities that my colleagues have found. >>  I >>>Imagine that, having an operating system that doesn't make you reboot. ' >>>Gee, where could I get one of those?  >>  K >> I don't think any vendor markets them :-)  I have some machines that run  >> aG >> "legacy" OS which no one talks about.  It's so old that it's only on 
 >> versionF >> 7.2 (7.3 is available) not 2000.  I really should "upgrade" to a PCI >> running this 2000 or newer XP so that I can waste as much time at work  >> and keep the I >> helpdesk in business.  I must learn to swear as well as my colleagues.  >>   >> Regards, Paddy  >>   >>  K > but, Paddy, those legacy apps won't run on XP so you'll have to hire some F > script kiddies to rewrite them so that they will run on 2000 or XP. I > Shoot, my voicemail system still runs on 6.2 and is nearly flawless. (I " > hate the NETBUI protocol though)  < Note, that was DOS 6.2.  But at least it was kind of stable.   --   Art Rice Tandem Admin Special Data Processing Corp ----------------------------* All opinions are my own and do not reflect* the views of the above mentioned employer.   ------------------------------  # Date: Mon, 29 Oct 2001 02:14:44 GMT * From: cjt & trefoil <cheljuba@prodigy.net>, Subject: Re: Windows XP reality check please+ Message-ID: <3BDCBBB8.4CB5406F@prodigy.net>    Paul Sture wrote:  >  <snip>S > I believe I saw the comment during the XP launch that it would save us an hour of N > lost time per week. 48 weeks x 5 years x hourly rate comes to a nice invoice	 > figure.  > E > Multiply by the number of PCs at work and it is indeed a large sum.  >  > ___  > Paul Sture
 > Switzerland   M So you would suggest the advertising slogan: "Windows XP -- it wastes less of % your time than our earlier products?"    ------------------------------   End of INFO-VAX 2001.601 ************************