1 INFO-VAX	Mon, 04 Dec 2000	Volume 2000 : Issue 676       Contents:. Re: AlphaStation 255/300 and large disk drives. Re: AlphaStation 255/300 and large disk drives. Re: AlphaStation 255/300 and large disk drives; Re: Compaq Secure Web Server for OpenVMS - access violation ; Re: Compaq Secure Web Server for OpenVMS - access violation ; Re: Compaq Secure Web Server for OpenVMS - access violation ; Re: Compaq Secure Web Server for OpenVMS - access violation   Re: OpenVMS 7.3 - The right dateE Trouble with isacfg on AXPpci33 (noname) and DE205 (isa network card)  Re: Two hosts sharing RA system  Re: Two hosts sharing RA system  Username length limits Re: Username length limits Re: VMS vs. Tru64 Unix Re: VMS vs. Tru64 Unix  F ----------------------------------------------------------------------  # Date: Sun, 03 Dec 2000 21:07:13 GMT ! From: "dls2" <dlshearer@home.com> 7 Subject: Re: AlphaStation 255/300 and large disk drives < Message-ID: <5CyW5.16570$w35.2972893@news1.rdc1.nj.home.com>  8 "David J. Dachtera" <djesys.nospam@earthlink.net> wrote: > Javier Henderson wrote:  > > J > >         I'm trying to find what's the size limit for disk drives on an& > > AlphaStation 255/300 with VMS 7.2. > ? > I'm not sure there is a limit, outside of hardware and driver C > limitations which seem quite liberal. 18+GB drives are not at all 6 > uncommon, and some people build EXTREMELY LARGE RAID? > sets and present them to Alpha/VMS systems with few problems.  > F > >         This is a hobby system, so buying genuine RZ series is out > > of the question. > < > Rather depends on wther you can live with "small" disks... > C > I found a guy on eBay selling RZ28B-VAs (circa 2GB) for $40 each. A > If you don't have a storage shelf, you can take them out of the ? > canisters and use them bare (may need SCSI id. jumpers). I've / > picked up a BA350 shelf for less than $200US.  > ? > Castlewood's Orb (cartridge drive) works reasonably well with A > OVMS, but only if you have a way to install the latest firmware B > update (needs a DOS machine with a SCSI card). The drive retails? > around $179.95US and the cartridges (circa 2.2GB unformatted, # > 2.05GBF/ODS) are around $29.95US.   = A brand new 9GB Seagate Barracuda costs "only" $250-300, and, - as a result, seems to be more cost effective.        --  Derrick Shearer    ------------------------------  # Date: Sun, 03 Dec 2000 21:51:52 GMT $ From: stevens@vtaks.com (D. Stevans)7 Subject: Re: AlphaStation 255/300 and large disk drives . Message-ID: <3a2ab846.48735800@news.vtacs.com>  A On 02 Dec 2000 22:04:59 -0800, Javier Henderson <javier@kjsl.com>  wrote:  ? >	This is a hobby system, so buying genuine RZ series is out of G >the question. And finding "small" 4GB drives locally has been a trying 
 >adventure...   9 	I wouldn't give up hope of finding RZ drives for a hobby D project.  Lots of flea markets sell them nowadays, dec stuff usuallyD shifts between expensive or cheap, since on occasion some indiviualsE get their hands on it, but dont know what it is or what to do with it  (i.e. not plug n play :P ). ? 	Three years ago I went to a flea market and saw a stack of two C SZ12 expansion boxes with TZ30 tape drives.  The person offered $10 D for each.  The one I picked had an RZ57 1 gig in it (back when 1 gigE was still big space).  It took some tinkering with scsi keys, since I F couldnt find the jumper info online...but eventually I figured out the6 proper sequence, and scsi ports 3 and 5 were filled :)  ; 	The box looked like it was dropped at one time (had a huge C dent in the side), so I felt very fortunate the drive still worked. ? It's usually left off due to power usage (sounds like a turbine D starting up), but it's worked very nice ever since.  Now I just need! to find some tapes for the tk-30.    Dan    ------------------------------  % Date: Sun, 03 Dec 2000 19:33:43 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> 7 Subject: Re: AlphaStation 255/300 and large disk drives - Message-ID: <3A2AF477.2AD1BA08@earthlink.net>    "D. Stevans" wrote:  > [snip] > Now I just need # > to find some tapes for the tk-30.    Uses TK50-K, just like TK50.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Sun, 03 Dec 2000 13:26:50 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> D Subject: Re: Compaq Secure Web Server for OpenVMS - access violation- Message-ID: <3A2A9E7A.F3B53530@earthlink.net>    Paul Sture wrote:  > 3 > In article <90bv0f$t83$1@nnrp1.deja.com>,  wrote:  > > From: jgessling@my-deja.com  > > Newsgroups: comp.os.vms H > > Subject: Re: Compaq Secure Web Server for OpenVMS - access violation' > > Date: Sat, 02 Dec 2000 23:01:36 GMT  > > / > > In article <VA.000001ac.1dceb595@sture.ch>, C > > > According to the Accounting utility, the final status code os 
 > > 10000024, G > > > text %SYSTEM-F-NOPRIV, isufficient privilege or object protection  > > > violation. > > >  > > D > > I'm not totaly sure, since I didn't work on it too much.  But myI > > experience was that the apache account needed lost of privs to really * > > work right.  That's what I did anyway. > > " > I'm not sure I like that idea... > H > But I've got it working on my home system with just NETMBX and TMPMBX.. > This is the CSWS version, dated 28-SEP-2000.  - Perhaps a larger difference is your IP stack.   A Both TCPware and Multinet need a patch to allow the actions which / require privilege in the unpatched IP software.   G Check Deja for relevant posts - I don't recall the exact details, but I E believe it's an issue regarding changing socket characteristics after . ownership of the socket has been relinquished.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Mon, 04 Dec 2000 01:39:57 +0100   From: Paul Sture <paul@sture.ch>D Subject: Re: Compaq Secure Web Server for OpenVMS - access violation+ Message-ID: <VA.000001b5.25c32235@sture.ch>   F In article <3A2A9E7A.F3B53530@earthlink.net>, David J. Dachtera wrote:9 > From: "David J. Dachtera" <djesys.nospam@earthlink.net>  > Newsgroups: comp.os.vms F > Subject: Re: Compaq Secure Web Server for OpenVMS - access violation' > Date: Sun, 03 Dec 2000 13:26:50 -0600  >  > Paul Sture wrote:  > > 5 > > In article <90bv0f$t83$1@nnrp1.deja.com>,  wrote: ! > > > From: jgessling@my-deja.com  > > > Newsgroups: comp.os.vms J > > > Subject: Re: Compaq Secure Web Server for OpenVMS - access violation) > > > Date: Sat, 02 Dec 2000 23:01:36 GMT  > > > 1 > > > In article <VA.000001ac.1dceb595@sture.ch>, E > > > > According to the Accounting utility, the final status code os  > > > 10000024, I > > > > text %SYSTEM-F-NOPRIV, isufficient privilege or object protection  > > > > violation. > > > >  > > > F > > > I'm not totaly sure, since I didn't work on it too much.  But myK > > > experience was that the apache account needed lost of privs to really , > > > work right.  That's what I did anyway. > > > $ > > I'm not sure I like that idea... > > J > > But I've got it working on my home system with just NETMBX and TMPMBX.0 > > This is the CSWS version, dated 28-SEP-2000. > / > Perhaps a larger difference is your IP stack.  >  Nup. Running TCPIP here.  K I know should try the others, but I use TCPIP at work, so it's a good idea  I (IMHO) to get practice at something which is useful during working hours.   C > Both TCPware and Multinet need a patch to allow the actions which 1 > require privilege in the unpatched IP software.  > I > Check Deja for relevant posts - I don't recall the exact details, but I G > believe it's an issue regarding changing socket characteristics after 0 > ownership of the socket has been relinquished. > D I struggle with Deja (doesn't everyone?). I've got most posts since I mid-April sitting here, less than a minute away, if I can remember a key  B phrase. Definitely faster, but I wish I had this facility at work. ___ 
 Paul Sture Switzerland    ------------------------------  % Date: Mon, 04 Dec 2000 02:07:54 +0100 B From: Michiel Erens <I.dont.want.spam@this.mailaddress.is.invalid>D Subject: Re: Compaq Secure Web Server for OpenVMS - access violation7 Message-ID: <3A2AEE6A.66EF@this.mailaddress.is.invalid>    Paul Sture wrote:  > H > In article <3A2A9E7A.F3B53530@earthlink.net>, David J. Dachtera wrote:; > > From: "David J. Dachtera" <djesys.nospam@earthlink.net>  > > K > > Check Deja for relevant posts - I don't recall the exact details, but I I > > believe it's an issue regarding changing socket characteristics after 2 > > ownership of the socket has been relinquished. > > E > I struggle with Deja (doesn't everyone?). I've got most posts since J > mid-April sitting here, less than a minute away, if I can remember a keyD > phrase. Definitely faster, but I wish I had this facility at work.  B It is so easy. Do a powersearch with the words apache and tcpware,9 and the forum comp.os.vms and you get 15 postings, one of  them from Hunter Goatley :  : http://x59.deja.com/=dnc/[ST_rn=ps]/getdoc.xp?AN=666856727  9 (the message id is <39b7a642.10211102@swen.process.com> )    --   ME   ------------------------------  % Date: Mon, 04 Dec 2000 01:01:48 -0500 " From: Dan Sugalski <dan@sidhe.org>D Subject: Re: Compaq Secure Web Server for OpenVMS - access violation: Message-ID: <5.0.1.4.0.20001204010057.01e22610@24.8.96.48>  , At 10:44 AM 12/3/00 +0100, Paul Sture wrote:2 >In article <90bv0f$t83$1@nnrp1.deja.com>,  wrote: > > From: jgessling@my-deja.com  > > Newsgroups: comp.os.vms H > > Subject: Re: Compaq Secure Web Server for OpenVMS - access violation' > > Date: Sat, 02 Dec 2000 23:01:36 GMT  > > / > > In article <VA.000001ac.1dceb595@sture.ch>, C > > > According to the Accounting utility, the final status code os 
 > > 10000024, G > > > text %SYSTEM-F-NOPRIV, isufficient privilege or object protection  > > > violation. > > >  > > D > > I'm not totaly sure, since I didn't work on it too much.  But myI > > experience was that the apache account needed lost of privs to really * > > work right.  That's what I did anyway. > > ! >I'm not sure I like that idea...  > G >But I've got it working on my home system with just NETMBX and TMPMBX. - >This is the CSWS version, dated 28-SEP-2000.  > C >The difference is that at home I have V7.2-1 rather than V7.2-1H1.   L That's not that big a difference. Are you sure you installed and started it L properly? It temporarily needs privs to bind port 80 and I think to install 6 an image with privs. The main server doesn't need 'em.   					Dan  I --------------------------------------"it's like this"------------------- 2 Dan Sugalski                          even samurai? dan@sidhe.org                         have teddy bears and even ;                                       teddy bears get drunk    ------------------------------  % Date: Sun, 03 Dec 2000 13:23:59 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> ) Subject: Re: OpenVMS 7.3 - The right date - Message-ID: <3A2A9DCF.CBEB895D@earthlink.net>    "Volckaerts, Tommy" wrote: > I > OpenVMS 7.3 wil be released in March 2001 followed by 7.3.1 (which will E > include updates that were not ready to be put in 7.3) in Fall 2001.  > % > Just had a training on the subject.   F Hhmmm... Sounds like candidates for the upgrade should wait 'til Q4 of4 CY2001 to make the next step along the upgrade path.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  $ Date: Sun, 3 Dec 2000 23:56:10 -0500& From: "Mike Burch" <mgburch@smart.net>N Subject: Trouble with isacfg on AXPpci33 (noname) and DE205 (isa network card)/ Message-ID: <t2m8ppld10pk2e@corp.supernews.com>   7 Well, I've got my noname system up and running OpenVMS, ( but now I want to put it on the network.= I got a supported isa network card, the DE205, and put it in. & But, I can't add it to the isa config." The srm tells me there is no room.> Both 'add_de205' and 'isacfg ... -mk ...' give the same error.  + If I type 'isacfg -all', I see six entries: . keyboard, mouse, com1, com2, lpt1, and floppy.1 That is all.  Surely, there is room for one more.    How can I up the table size? Or is it fixed by the firmware? : If I reflash firmware with card installed would that help?7 OK, I'm grasping at straws here but I'm at my wits end. K Hopefully, its something simple I'm missing, but I don't see anthing in the  manual.    TIA for any suggestions, Mike   ------------------------------  $ Date: Mon, 4 Dec 2000 01:04:55 +01000 From: "Taki Pioy" <nospam@localhost.spamwarn.zz>( Subject: Re: Two hosts sharing RA system4 Message-ID: <4fBW5.2551$tm.8497@nntpserver.swip.net>  B "David J. Dachtera" <djesys.nospam@earthlink.net> wrote in message' news:3A2A8B40.FEEE484D@earthlink.net...  > Taki Pioy wrote: > > C > > Has anyone managed to configure and run sharing a RA450 storage B > > system, with one HSZ50 controller, between two Alphaserver/VMSG > > non-clustered hosts? I'm going to use this during an "upgrade" only : > > for a month or two, but still for production purposes. > (...)  > B > There was another thread this past week (I believe) where it wasE > mentioned that some SCSI controllers used in DS20s (I think) do not F > support SCSI clustering. Beyond that, I don't think shared SCSI is a% > real good idea, but that's just me.  > 	 > FWIW...  >   B Well, I know that the configuration is not supported, at least not without Cluster license.  @ And I'm not too comfortable about the idea either. But with only< one raid array system and the data access requirements (e.g.= "almost online" access to production data), it seems like the # proportionally right thing to do...   = Using an extra disk shelf and abuse the hot-swap capabilities  seems like a worse idea.   Regards  /Fad> (these are of course my own, possibly flawed, interpretations.A hopefully someone will point out and correct any mistakes herein)    ------------------------------  % Date: Sun, 03 Dec 2000 19:36:02 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> ( Subject: Re: Two hosts sharing RA system, Message-ID: <3A2AF502.8BEDC98@earthlink.net>   Taki Pioy wrote: > D > "David J. Dachtera" <djesys.nospam@earthlink.net> wrote in message) > news:3A2A8B40.FEEE484D@earthlink.net...  > > Taki Pioy wrote: > > > E > > > Has anyone managed to configure and run sharing a RA450 storage D > > > system, with one HSZ50 controller, between two Alphaserver/VMSI > > > non-clustered hosts? I'm going to use this during an "upgrade" only < > > > for a month or two, but still for production purposes.	 > > (...)  > > D > > There was another thread this past week (I believe) where it wasG > > mentioned that some SCSI controllers used in DS20s (I think) do not H > > support SCSI clustering. Beyond that, I don't think shared SCSI is a' > > real good idea, but that's just me.  > >  > > FWIW...  > >  > D > Well, I know that the configuration is not supported, at least not! > without Cluster license. [snip]   H I think it's more question of data safety rather than a licensing issue.  G In this case, however, I believe the issue is that it just flat doesn't 1 work (yeah - THAT definition of "not supported").a   -- i David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/a  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.e   ------------------------------   Date: 3 Dec 2000 16:36:00 -0500s9 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)n Subject: Username length limits-+ Message-ID: <0S1woE8pkhBY@eisner.decus.org>   e In article <DTiotGxQ0bj6-pn2-ztUIfd9rpkjC@localhost>, djweath@attglobal.net (Dave Weatherall) writes: ? > On Sat, 2 Dec 2000 00:03:09, Ray <lists@aik.tec.sc.us> wrote:a >  >> > s( >> > VMS or Unix?  Thanks for your help! >> > e >> AN >> One thing to consider is that Tru64 has an 8 character username limitation. > E > Only people with names longer than 8 characters really notice that t
 > problem :-)a  D Several of us have asked that VMS be forever capped at 12 characters+ for the benefit of Brian Schenkenberger :-)i   ------------------------------  # Date: Sun, 03 Dec 2000 22:37:14 GMTs= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) # Subject: Re: Username length limitsa0 Message-ID: <009F40D8.232131B9@SendSpamHere.ORG>  g In article <0S1woE8pkhBY@eisner.decus.org>, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes: f >In article <DTiotGxQ0bj6-pn2-ztUIfd9rpkjC@localhost>, djweath@attglobal.net (Dave Weatherall) writes:@ >> On Sat, 2 Dec 2000 00:03:09, Ray <lists@aik.tec.sc.us> wrote: >> 7 >>> > ) >>> > VMS or Unix?  Thanks for your help!- >>> >  >>> O >>> One thing to consider is that Tru64 has an 8 character username limitation.@ >> .F >> Only people with names longer than 8 characters really notice that  >> problem :-) >0E >Several of us have asked that VMS be forever capped at 12 characters,, >for the benefit of Brian Schenkenberger :-)  F For those folks counting in Flori-duh, that would be enough characters  to handle my entire surname.  ;)   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM1             O city, n., 1. a place where trees are cut down and streets are named after them.t   ------------------------------   Date: 4 Dec 2000 00:31 -0400 From: hein@eps.ko.dec.c*me Subject: Re: VMS vs. Tru64 Unix % Message-ID: <4DEC200000311959@miasys>'  I In article <9092lg$qr4$1@nnrp1.deja.com>, waybright@my-deja.com writes...eB >I should be more specific with my questions...I do appreciate the >detailed answers I've seen.  2 Caveat emperor.... free advice, worth every penny!= I though that other ereply was very appropriate: Are you sure > you want to take well intended (we hope) input from 'Joe Blow'> in a newsgroup to take a big decision? Be sure to ask (Compaq) for formal advice / details.  
 >Performance:  > E >Galaxy, I understand, overcomes the CPU overhead issue with VMS SMP. C >The sweet spot for VMS was 12 CPU's and anything over that you hade   More like 8 really...   H >diminishing returns, or it did not scale linearly.  With Galaxy you canF >continue to scale linearly as you add CPU's to take full advantage ofM >the GS platform.  Question:  Are there other performance benefits of Galaxy?   D The GS platform is NUMA: 'Near' Uniform Memory Access, so it is more; costly (performance) to access remote memory versus local. a7 Galaxy readily allows contro over remote/global memory.V  N >Also, does UNIX have the same CPU scaling issue that VMS had prior to Galaxy?  F Less so. VMS suffers from primary CPU overload, and low grained systemH locks (IOLOCK8). VMS has fixed much of this and continues to improve butD IMHO Unix still has the edge for high SMP count for may application.K Unix 5.1 can/will do 'distributed IO' all the time, not just for 'fastpath'0 devices.    F >Compaq/Oracle recently announced the highest TPM-C benchmark ever for >an Oracle DB using Tru64 UNIX.   D And consequently Oracle on Unix has seen the first Numa enhancement ; and they have the 'spike' optimiations available and so on.oD VMS is sure to follow closely for many optimiaitons found during theD benchmark preparations, but as much as I like VMS, IMHO Unix has and@ is likely to retain the 'absolute' performance lead. Still, the @ performance is just one of the many deciding factors. As long asI you can buy a box big enough to do the job, other factor may matter more.o  * >  Is there a similar benchmark for VMS or) >does anyone have some performance stats?1  J No formal data, and informal data is 'cooked' (sometimes unknowingly) moreE often than not. As Hoff suggests... can you use your own application?s    H >Are there any whitepapers that identify Tru64 UNIX position in the UNIXA >market?  Is Tru64 UNIX considered number one in the UNIX market?c  J IMHO is is number one in QUALITY. I know is is not number one in quantity.K At various time it is/was the fastests growing Unix, if that says anything. E There are lots of positioning whitepapers. Start at the Web-pages, orf% ask a local Compaq Unix 'ambassador'.r# 	- http://www.tru64unix.compaq.com/h) 	- http://tru64unix.compaq.com/idcrel.doc'T 	- http://www.compaq.com/newsroom/pr/2000/0,1494,wp%7E14583_2!ob%7E27972_1_1,00.html    ) Anyway, Just go VMS + RDB. It's the best.,  
 Good Luck! 	 Hein.    ------------------------------   Date: 4 Dec 2000 00:43 -0400 From: hein@eps.ko.dec.c*m. Subject: Re: VMS vs. Tru64 Unixs% Message-ID: <4DEC200000432293@miasys>5  d In article <87pujbpiyq.fsf@k9.prep.synonet.com>, Paul Repacholi <prep@k9.prep.synonet.com> writes... >waybright@my-deja.com writes: >  >>  H >> Also, does UNIX have the same CPU scaling issue that VMS had prior to
 >> Galaxy? > ; >Worse. VMS does IO by DMA to the process. You only contendp; >with processes that are sharing the file. Unix does IO viad? >the kernel buffer cache. you are contending with ALL processesr= >IO ( all of it ) and memory requirements. The contention cani >get pretty horrid real fast.t  A Well, As usual it depends... VMS has/will have (depending on whatdA version you run) it's XFC and thus will have the same challenges.dE Unix has RADicalied/NUMAtied its UBC. For 'absolute' performance Unix E applcation implementors use 'RAW' IO also avoiding a shared (central)iH buffer. Unix (currently) has the additional benefit of using a RAD-local3 adapter for it's IO if an access path is available.a  = >Also, tests I've run have shown 20:1 advantage to VMS vs DU4t; >in heavy paging. This was with FFTs, so don't expect it to  >at all repressent a DB case.   3 Production system are not supposed to page or swap.eK No matter which one is better or worse, Test showing a magnitude differenceaH in end user performance on the same box are 'suspect' in the very least.F Either they uncovered a anomaly which is probably fixable or they just8 show un-equal tuning/coding/setup skills for the tester.  J As much as I like VMS, the 20x performance differences are more often thanL not in the favor of Unix, often involving careless file opens/writes becauseB out-of-the-box, VMS is in failsafe/paranoia/minimie resource mode.4 This is goodness to many, but off-putting to others.   Cheers,  	Hein.   ------------------------------   End of INFO-VAX 2000.676 ************************