1 INFO-VAX	Mon, 26 May 2003	Volume 2003 : Issue 290       Contents:G Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...") G Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...") G Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...") ' Advanced Server: External Authorization + Re: Advanced Server: External Authorization + Re: Advanced Server: External Authorization + Re: Advanced Server: External Authorization % Re: advsyscon remote shadow on alpha?  Re: Help maping RDB   Re: Johnny English is a VMS user* Re: New [to me] error showing in errorlog. Re: Portents of VMS death  Re: Portents of VMS death  Re: Portents of VMS death  Re: Samba for OpenVMS How-To?  Re: Samba for OpenVMS How-To?  Re: Samba for OpenVMS How-To?  Re: Samba for OpenVMS How-To? $ SLS + DCSC + ACSLS: Spanning volumes  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue  Re: SYS$QIOW - Performance Issue2 Re: ZLE data synchronise with different line speed  F ----------------------------------------------------------------------  # Date: Mon, 26 May 2003 13:47:16 GMT # From: "John Smith" <a@nonymous.com> P Subject: Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...")J Message-ID: <EVoAa.271812$w7k.252193@news04.bloor.is.net.cable.rogers.com>  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message # news:3ED0FDEA.3968D90A@istop.com... B > > > And don't forget that to be listed as an approved vendor for many > > > categories< > > > of customer, ie. gov't, .edu, and even some commercial
 customers,
 > > > they'll B > > > want to see some business history, financials, and god knows what > > > else.  > @ > In this age of IT becoming commodity, is that really necessary	 anymore ? D > Wouldn't that be a case of the VMS sales infrastructure not having evolved  > with the times ?    < VMS isn't a commodity, which is both a blessing and a curse.  F Your customers will want to know that you have some staying power when( the product you sell isn't a Billy-box..   ------------------------------  % Date: Mon, 26 May 2003 11:12:32 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> P Subject: Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...")' Message-ID: <3ED23CF0.890D904C@fsi.net>    John Smith wrote:  > 9 > "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message % > news:3ED0FDEA.3968D90A@istop.com... D > > > > And don't forget that to be listed as an approved vendor for > many > > > > categories> > > > > of customer, ie. gov't, .edu, and even some commercial > customers, > > > > they'll D > > > > want to see some business history, financials, and god knows > what
 > > > > else.  > > B > > In this age of IT becoming commodity, is that really necessary > anymore ? F > > Wouldn't that be a case of the VMS sales infrastructure not having	 > evolved  > > with the times ? > > > VMS isn't a commodity, which is both a blessing and a curse. > H > Your customers will want to know that you have some staying power when* > the product you sell isn't a Billy-box..  7 True. VMS's history is both a blessing and a challenge.   D It survived the attack of GQ Palmer, it survived the incompetence ofF Compaq, and it shows every sign of being able to survive the apathy of HP.   6 Tough to sell; but, bring in the spin experts and, ...   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Mon, 26 May 2003 17:14:17 GMT # From: "John Smith" <a@nonymous.com> P Subject: Re: (Very) Affordable VMS on older gear (was: "VMS will be around ...")J Message-ID: <JXrAa.272523$w7k.165666@news04.bloor.is.net.cable.rogers.com>  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3ED23CF0.890D904C@fsi.net...  > John Smith wrote:  > > ; > > "JF Mezei" <jfmezei.spamnot@istop.com> wrote in message ' > > news:3ED0FDEA.3968D90A@istop.com... F > > > > > And don't forget that to be listed as an approved vendor for > > many > > > > > categories@ > > > > > of customer, ie. gov't, .edu, and even some commercial > > customers, > > > > > they'll F > > > > > want to see some business history, financials, and god knows > > what > > > > > else.  > > > D > > > In this age of IT becoming commodity, is that really necessary
 > > anymore ? A > > > Wouldn't that be a case of the VMS sales infrastructure not  having > > evolved  > > > with the times ? > > @ > > VMS isn't a commodity, which is both a blessing and a curse. > > E > > Your customers will want to know that you have some staying power  when, > > the product you sell isn't a Billy-box.. > 9 > True. VMS's history is both a blessing and a challenge.  > F > It survived the attack of GQ Palmer, it survived the incompetence ofE > Compaq, and it shows every sign of being able to survive the apathy  of > HP.  > 8 > Tough to sell; but, bring in the spin experts and, ...    C No need to 'spin' the product. All you have to do is make sure that E the customer's mind is open and the features and benefits of VMS will 	 register.   C "You're a reasonable man/woman. I can see that. I can tell that you A carefully judge situations based on the facts and would chose the E solution that provides the best performance and availability for your C company's needs. You know all too well the cost of downtime and how B that can adversely affect the corporation, not to mention your own@ career prospects. That's why I'm certain after you carefully and= objectively compare what we are showing you and all the other F environments available to your company, you'll choose OpenVMS for this
 application."   E The only hard part is convincing them that HP isn't going to kill it, B and the best way to do that is to have carly (tm) make some highlyF visible non-ambiguous announcement to that effect.....she'd be wise toB run the wording of that pronouncement through c.o.v first via Sue.   ------------------------------  % Date: Mon, 26 May 2003 13:49:51 +0300 - From: "Kari Keronen" <kari.keronen@digita.fi> 0 Subject: Advanced Server: External Authorization; Message-ID: <ojmAa.7609$M8.5513477@reader1.news.jippii.net>   G Is it possible to configure Advanced Server 7.3A/OpenVMS 7.3-1 so, that G changing password in Windows Domain (PDC Win 2000) will change password K automatically also in VMS system ? I have looked manuals, but don't seem to J find enough information. If someone has configured this I'd be grateful to	 know how.    -Kari-   ------------------------------    Date: 26 May 2003 09:34:38 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 4 Subject: Re: Advanced Server: External Authorization3 Message-ID: <8+PDe2X27wsj@eisner.encompasserve.org>   k In article <ojmAa.7609$M8.5513477@reader1.news.jippii.net>, "Kari Keronen" <kari.keronen@digita.fi> writes: I > Is it possible to configure Advanced Server 7.3A/OpenVMS 7.3-1 so, that I > changing password in Windows Domain (PDC Win 2000) will change password $ > automatically also in VMS system ?  F Windows provides no outbound messaging of that sort, but with a normalI configuration on VMS the first login to VMS after the change will consult J the Windows system and allow or prohibit access based on the new password.   ------------------------------   Date: 26 May 2003 16:25:26 GMT* From: Terry Aardema <taardema@nrcan.gc.ca>4 Subject: Re: Advanced Server: External Authorization< Message-ID: <Xns93876A09C52E5taardemanrcangcca@132.156.36.9>  0 Kilgallen@SpamCop.net (Larry Kilgallen) wrote in, news:8+PDe2X27wsj@eisner.encompasserve.org:   C > In article <ojmAa.7609$M8.5513477@reader1.news.jippii.net>, "Kari , > Keronen" <kari.keronen@digita.fi> writes: E >> Is it possible to configure Advanced Server 7.3A/OpenVMS 7.3-1 so, F >> that changing password in Windows Domain (PDC Win 2000) will change. >> password automatically also in VMS system ? > H > Windows provides no outbound messaging of that sort, but with a normalC > configuration on VMS the first login to VMS after the change will F > consult the Windows system and allow or prohibit access based on the > new password.  >   K Correct. A successful login will also set (syncronize) the VMS password to  F the Windows password (allowing for case folding) *if and ONLY if* the L Windows password contains only charcaters that are legal in a VMS password. L If the Windows password is illegal under VMS, the login will still succeed, * but no password syncronization will occur.  L For normal interactive logins this is not a concern; the password is always G validated by the Windows domain. A lack of password syncronization can  K affect other types of logins, such as via FTP (for older UCX/TCPIP stacks;  J I know the current (V5.3) TCPIP actually uses the domain password and not L the local password, but I don't remember when this was brought in. I do not C know about MULTINET et. al.), DECWindows and a whole host of other  K applications. For those applications that need your password but only look  E in the UAF, you need to remember to log onto VMS after changing your  I Windows password, and to only choose Windows passwords that are legal in  C VMS, or you'll be in a world where nothing ever works consistently.   
 Terry Aardema    ------------------------------    Date: 26 May 2003 11:58:59 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) 4 Subject: Re: Advanced Server: External Authorization3 Message-ID: <W+vFEIFIgS9f@eisner.encompasserve.org>   i In article <Xns93876A09C52E5taardemanrcangcca@132.156.36.9>, Terry Aardema <taardema@nrcan.gc.ca> writes:   M > applications. For those applications that need your password but only look  G > in the UAF, you need to remember to log onto VMS after changing your  K > Windows password, and to only choose Windows passwords that are legal in  E > VMS, or you'll be in a world where nothing ever works consistently.   ? Any such applications on Alpha should be changed to use the new > $ACM system service effective with VMS V7.3-1.  Then they will< get not only the External Authentication connection but also< any changes in handling breakin evasion, etc.  Some internal= VMS and layered product components had to change over several : generations of ancillary security checks -- the new system7 service should insulate applications from that problem.    ------------------------------  # Date: Mon, 26 May 2003 12:51:47 GMT " From:   VAXman-  @SendSpamHere.ORG. Subject: Re: advsyscon remote shadow on alpha?0 Message-ID: <00A206EC.83051444@SendSpamHere.ORG>  b In article <kuff-C4933A.10151925052003@news-east.giganews.com>, Hal Kuff <kuff@tessco.com> writes:1 >In article <00A2054D.E8B0450E@SendSpamHere.ORG>, # > VAXman-  @SendSpamHere.ORG wrote:  > L >> In article <kuff-C63589.05122824052003@news-east.giganews.com>, Hal Kuff  >> <kuff@tessco.com> writes:5 >> >Is anyone using advsyscon remote shadow on alpha?  >>  % >> Only for possible data corruption.  >>   >> -- : >> VAXman- OpenVMS APE certification number: AAA-0001      >> VAXman(at)TMESIS(dot)COM  >>            8 >>   "Well my son, life is like a beanstalk, isn't it?"  >  > D >Can I assume then that the huge license fee these guys want is not D >inline with their performance... I have heard much the same thing.  > A >Their sales dude claims that post September 11th they saved the   >financial industry.  A ... and remember that Al Gore invented the internet and that Bill  Gates invented the computer.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------  # Date: Mon, 26 May 2003 17:00:46 GMT # From: "John Smith" <a@nonymous.com>  Subject: Re: Help maping RDBI Message-ID: <2LrAa.272500$w7k.21372@news04.bloor.is.net.cable.rogers.com>   0 "bayden cline" <bayden@isys.ca> wrote in messageC news:AIaAa.262748$w7k.23230@news04.bloor.is.net.cable.rogers.com... C > Hello, i am trying to figure out a way to map out how the various 	 tables in C > an RDB 5.1 database are linked together, just wondering if anyone  here has7 > any sujestions for how to do this.  Thanks in advance     E V3.x of ERwin (now owned by Computer Associates) could connect to RDB C and read/reverse-engineer the schema. Not sure if v4.x has Rdb as a 4 'native' db any longer...if not, ODBC ought to work.   ------------------------------  # Date: Mon, 26 May 2003 17:39:16 GMT # From: "John Smith" <a@nonymous.com> ) Subject: Re: Johnny English is a VMS user J Message-ID: <8jsAa.272576$w7k.134096@news04.bloor.is.net.cable.rogers.com>  A "Doc.Cypher" <Use-Author-Address-Header@[127.1]> wrote in message 0 news:20030501093534.15384.qmail@gacracker.org...7 > NOTE: This message was sent thru a mail2news gateway. : > No effort was made to verify the identity of the sender.: > -------------------------------------------------------- > 9 > On 1 May 03, p_sture@elias.decus.ch (Paul Sture) wrote: 6 > >In article <3EAF89C8.B6D08864@127.0.0.1>, Nic Clews# > ><sendspamhere@127.0.0.1> writes:  >  > >> G.A.R.T.N.E.R.  > >> > >> (Personal opinion)  > > D > >And a box of Swiss chocolates to the person who comes up with the mostC > >apt/most humourous rendition of what the initials G.A.R.T.N.E.R.  stand 
 > >for :-) > # > OK, a one cup of coffee effort...  >  >  G enerate
 >  A nnual >  R evenue  >  T elling  >  N onsense to 	 >  E ager  >  R eaders     	 Good one.   ; In the Bond movies we were exposed to an evil organization: C SPECTRE - the Special Executive for Counterintelligence, Terrorism,  Revenge and Extortion   F which had its genesis in the mind of Ian Fleming, who was a WWII navalE intelligence officer before he wrote the James Bond series of novels. @ SPECTRE is based in part on a real-life analog known to Fleming, SMERSH.   A The term SMERSH comes from the Russian words "smert' shpionam" or B "Death to Spies!". SMERSH was Soviet military counter-intelligenceD during the later part of WWII. It was officially completely separate/ from the security organs; the NKVD and the KGB.     E As it turns out, IT has it's own nefarious organization to deal with:  GARTNER   B Just as evil in its own way - a great proponent of disinformation, fear, uncertainty, and doubt.    ------------------------------  % Date: Mon, 26 May 2003 11:25:26 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)3 Subject: Re: New [to me] error showing in errorlog. L Message-ID: <rdeininger-2605031125270001@user-105n92m.dialup.mindspring.com>  O In article <qK%za.10583$Kn1.143045@news.chello.at>, peter@langstoeger.at wrote:      >Exactly. Or in other words:I >Why is this that one can no longer look at own errors with pure VMS ????   I As I understand it, DECevent and Compaq Analyze originated in the service J organization, not VMS Engineering.  Both tools attempt to support multiple OS's.      >First it was pure VMS. N >Then you needed DECevent (downloadable, show err = w/o PAK, analyze = w/ PAK)H >Later you needed COMPAQ Analyze (CD with h/w buy and what else, PAKs ?)  ' I think Compaq Analyze is downloadable.    >Now (at HPQ) you need what ?   6 Still Compaq Analyze, but I believe it has a new name.  O >And don't confuse it with the products to connect to the service centers (like P >DSNlink, DIA ?, WIS ?, ...) or the product collections (like AES, WEBES ?, ...)  H It's too late!  I've been confused by this quagmire of tools an acronyms for years...   ------------------------------    Date: 26 May 2003 02:54:58 -0700. From: mistdragon@zdnetonebox.com (mist dragon)" Subject: Re: Portents of VMS death= Message-ID: <7500353b.0305260154.7fd820c9@posting.google.com>    But as an OS, I don't see N > why you call it monolithic. It supports Unix's modern GUI (x windows) and isK > compatible with HP-UX's Motif (although VMS's version of motif is older).   D I refer to OS scheduling and kernel being monolithic. The processors< are not shared equally between processors serving kernel andB application systems. The same applies to many more older operatingC systems where the system performance freezes when most of operating D system load is using the same cpu's not able to take load from otherB cpu's that may have slices left. Kernel isnt just granular enough.> Also the kernel modules are not dynamic - you cant easily takeD subsystems off and replace them with other ones that may be the case@ if your system is not doing the thing that it is designed to do.  K > But VMS has the potential to not only keep up with the Jones, but in many L > aspects, be the leader of the pack.  Scalability is still a VMS advantage,, > which is not being leveraged by Compaq/HP.  A I question VMS scalability. Referring to my earlier example - for F example io subsystem and tcp/ip are far more scalable on tru64 than inC vms using the same large alpha. Vms freezes around 5000 and on that @ load tru64 has only used about 3% of resources. All the possibleE methods and best gurus were used to try to tune vms but in the end it  could not just perform enough).   C I believe the reason for old operating systems going away is not of F how good they are, but how much software they have. Unixes have almostE common api (almost=excluding some real system calls rarely used - Aix F is a bit more different than others, but only on user tools level) andD they are optimized the same way (in vms for example, a call and forkA consume lot of cycles, whereas in unix these are very light. This D leads to software where the individual sections of code are large inD vms and small in unix and porting of optimized code is difficult - IB used to do porting where the only target was to make it fast to myE living when vms and unix were both strong and did a lot of vms sw :).   C Before linux came along all systems were fighting off microsoft and @ now they are tuning to linux. Some do it because linux really IS* better and some because its not microsoft.  C In these days you basically have two server operating systems (unix F variances and derivates and microsofts os). The fact also is no-one inC their right mind would put datacenter to microsoft os. It cant just E take the load and the kernel is not granular enough. 2 CPU system can E share quite nicely, but 3rd processor gets around 20% and 4th 13%. If E you go beyong that you have to either work on clusters with os images E of 4 or purchase sw from suppliers that have their own knifed HAL but C then you have the same problem with sw that becomes wery customized D too. Microsofts load balancing is a laugh and the overall quality isF horrible: memory leaks cause the places I know to reboot their serversE every night whereas unix variants are usually up for years at a time.   F As for the other discussion of whether microsoft nt would be VMS basedE I meant the design, not the actual code :) Sorry for confusion. David D Cutler took (according his own words in inside windows nt) the ideasD of how to build the next generation vms when dec cut the funding andD built nt on it. According the human nature and they way projects areE built i wouldnt be a minute of surprised if he did reuse the ideas he @ had (not code) what he had insteading inventing nt from scratch.  ? Of course we may debate what killed vms, but I belive it was 1) E protectionism, 2) expensiveness 3) nih-culture and finally: glassroom F only. Market has moved from datacenter to desktop and now from desktopF to datacenter. On the way the ones that could not move to desktop diedF and now the ones that cannot live in glassroom will die. The ones thatB originally lost it (like vms) are not gonna live in the long term.   Mist   ------------------------------  % Date: Mon, 26 May 2003 08:41:04 -0400 + From: "Ron Milen" <milenronald@mailbag.net> " Subject: Re: Portents of VMS death! Message-ID: <3ed20a39_1@newsfeed>   ; "mist dragon" <mistdragon@zdnetonebox.com> wrote in message 7 news:7500353b.0305260154.7fd820c9@posting.google.com...  > But as an OS, I don't see I > > why you call it monolithic. It supports Unix's modern GUI (x windows)  and isE > > compatible with HP-UX's Motif (although VMS's version of motif is  older).  > F > I refer to OS scheduling and kernel being monolithic. The processors> > are not shared equally between processors serving kernel andD > application systems. The same applies to many more older operatingE > systems where the system performance freezes when most of operating F > system load is using the same cpu's not able to take load from otherD > cpu's that may have slices left. Kernel isnt just granular enough.@ > Also the kernel modules are not dynamic - you cant easily takeF > subsystems off and replace them with other ones that may be the caseB > if your system is not doing the thing that it is designed to do. >   H At the last HP user's group meeting in Cleveland they said that this wasG going to be changed in a future release, which I think was 8.2 (I won't J swear to it however).  If you want a contact at HP please e-mail me & I'll# send you the name & e-mail address.   	 Ron Milen  milenronald@yahoo.com    ------------------------------  % Date: Mon, 26 May 2003 11:18:53 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)" Subject: Re: Portents of VMS deathL Message-ID: <rdeininger-2605031118540001@user-105n92m.dialup.mindspring.com>  = In article <7500353b.0305260154.7fd820c9@posting.google.com>, / mistdragon@zdnetonebox.com (mist dragon) wrote:   7 >I refer to OS scheduling and kernel being monolithic.     >The processors = >are not shared equally between processors serving kernel and  >application systems.   G There are few subsystems in the VMS kernel that must run on the primary H processor.  Most of the kernel can run on any available processor.  WithE each recent release, less code is dependent on the primary processor.   . >The same applies to many more older operatingD >systems where the system performance freezes when most of operatingE >system load is using the same cpu's not able to take load from other C >cpu's that may have slices left. Kernel isnt just granular enough.   @ I've rarely seen VMS in this situation.  It typically requires a@ pathalogical workload, or a nincompoop at the controls, or both.  ? >Also the kernel modules are not dynamic - you cant easily take E >subsystems off and replace them with other ones that may be the case A >if your system is not doing the thing that it is designed to do.   J True up to a point.  VMS supports loading and unloading execlets while the> system is up.  But most of the current modules don't have thisH functionality included.  On the other hand, many parts of the kernel canF be tuned and configured on a live system, so there is seldom a need to reload execlets.   >I question VMS scalability.    H You question all sorts of things about VMS, but with little credibility.  & >Referring to my earlier example - forG >example io subsystem and tcp/ip are far more scalable on tru64 than in   >vms using the same large alpha.  H This is a pretty meaningless claim unless you specify which large alpha,I which version of each OS, and define the workload.  Cases are known where = VMS does better, and cases are known where Tru64 does better.     $ >Vms freezes around 5000 and on that1 >load tru64 has only used about 3% of resources.    I Well, I think a nincompoop must have set up "that load", whatever it is.  @ It isn't unusual to see VMS lag Tru64 by 10-15% on CPU-intensiveJ workloads, with reasonable tuning.  I've never seen such a wide gap as you	 describe.   H And VMS "freezes"?  I have (rarely) seen bugs lead to such symptoms, but; almost always the cause has been really, really bad tuning.   C I've seen VMS scale very well up to 16 processors, and occasionally G beyond.  But not on poorly-scaling hardware like the previous GS-series  systems.     >All the possible F >methods and best gurus were used to try to tune vms but in the end it  >could not just perform enough).  G Since you don't name names, we'll have to take your word that they were 7 the "best gurus".  Or perhaps you're just making it up.   D >I believe the reason for old operating systems going away is not ofG >how good they are, but how much software they have. Unixes have almost F >common api (almost=excluding some real system calls rarely used - AixG >is a bit more different than others, but only on user tools level) and E >they are optimized the same way (in vms for example, a call and fork > >consume lot of cycles, whereas in unix these are very light.   B Calls are rather light on VMS.  Fork() is a well-known performanceD problem, and anyone with a clue chooses an algorithm that uses it as little as possbile.    >This E >leads to software where the individual sections of code are large in A >vms and small in unix and porting of optimized code is difficult   8 I begin to wonder if we're talking about the same VMS...   ------------------------------  # Date: Mon, 26 May 2003 10:27:11 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton)& Subject: Re: Samba for OpenVMS How-To?. Message-ID: <3_lAa.987007$F1.119878@sccrnsc04>  [ In article <3ED17A8F.6C87DF5F@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:  >Jan-Erik Sderholm wrote: >>  ? >> Why not just subscribe to the SAMBA-VMS mailing list and ask " >> your specific questions there ? >>   >> See:  >>  , >> http://lists.samba.org/listinfo/samba-vms > 7 >Is there a place to easily find a digest of that list?   M David, I know of a searchable archive of that list, if that would be helpful:   1 http://marc.theaimsgroup.com/?l=samba-vms&r=1&w=2   K The archive goes back to 1998; I found it useful when installing Samba with  TCPware as the IP stack.   <snip>: >Now, I'm trying to generate the documentation to make theF >kit as usable as possible out of the "box", with the ultimate goal ofH >having the software be easily configurable after install time (yes, I'm >thinking PCSI here).    The distribution at:  + http://www.pi-net.dyndns.org/anonymous/jyc/   N includes a "step-by-step" at the bottom of the page for a compile/link/installL sequence; could that not be modified to fit within your envisioned PCSI kit?  K Configuration after install time became somewhat easier for me by using the N SWAT (Samba Web Administration Tool) included as part of the Samba package; itN contains good documentation/explanation of the Samba "man" page, as well as anO "express" option to quickly base-configure Samba.  Instructions/recommendations C on using the SWAT could be included in the PCSI kit, as part of the ! "post-installation" instructions.    <snip> >--  >David J. Dachtera >dba DJE Systems >http://www.djesys.com/  > ) >Unofficial Affordable OpenVMS Home Page:   >http://www.djesys.com/vms/soho/  A _________________________________________________________________ 0 Bradford J. Hamilton			"All opinions are my own"/ bMradAhamiPltSon@atMtAbi.cPoSm		"Lose the MAPS"    ------------------------------  % Date: Mon, 26 May 2003 11:23:56 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> & Subject: Re: Samba for OpenVMS How-To?& Message-ID: <3ED23F9C.F2BBC57@fsi.net>   "Bradford J. Hamilton" wrote:  > ] > In article <3ED17A8F.6C87DF5F@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:  > >Jan-Erik Sderholm wrote: > >>A > >> Why not just subscribe to the SAMBA-VMS mailing list and ask $ > >> your specific questions there ? > >>	 > >> See:  > >>. > >> http://lists.samba.org/listinfo/samba-vms > > 9 > >Is there a place to easily find a digest of that list?  > O > David, I know of a searchable archive of that list, if that would be helpful:  > 3 > http://marc.theaimsgroup.com/?l=samba-vms&r=1&w=2   H I found that one, also. Also found gname.org which provides a newsreader interface to the list.  M > The archive goes back to 1998; I found it useful when installing Samba with  > TCPware as the IP stack.  F I'm starting with Multinet just because that's what I've used the most; and it's what I have on my Alpha and MivroVAX here at home.    > <snip>< > >Now, I'm trying to generate the documentation to make theH > >kit as usable as possible out of the "box", with the ultimate goal ofJ > >having the software be easily configurable after install time (yes, I'm > >thinking PCSI here).  >  > The distribution at: > - > http://www.pi-net.dyndns.org/anonymous/jyc/  > P > includes a "step-by-step" at the bottom of the page for a compile/link/installN > sequence; could that not be modified to fit within your envisioned PCSI kit?  = That's exactly what I've done. I have a DCL proc. which edits @ INSTALL.COM and some others as supplied in JYC's distro. The kitF resulting from that installs directly from the CD without copying down the entire source tree.   H That's good because both VAX and Alpha are served by single distro. ThatC might not be possible or practical when/if I get to the PCSI stage. 
 Dunno yet.  M > Configuration after install time became somewhat easier for me by using the P > SWAT (Samba Web Administration Tool) included as part of the Samba package; itP > contains good documentation/explanation of the Samba "man" page, as well as anQ > "express" option to quickly base-configure Samba.  Instructions/recommendations E > on using the SWAT could be included in the PCSI kit, as part of the # > "post-installation" instructions.   F I haven't gotten as far as SWAT yet. It kinda flies in the face of oneB of my secondary goals to make the kit as "VMS" as possible withoutE requiring a lot of ancillary and/or non-VMS stuff to configure/manage B it. I was thinking of a DCL proc. ala NET$CONFIGURE that could ask/ questions and re-write SMB.CONF as appropriate.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Mon, 26 May 2003 11:47:44 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> & Subject: Re: Samba for OpenVMS How-To?' Message-ID: <3ED24530.ADC5FCAC@fsi.net>    "David J. Dachtera" wrote: >  > "Bradford J. Hamilton" wrote: 
 > > [snip]Q > > David, I know of a searchable archive of that list, if that would be helpful:  > > 5 > > http://marc.theaimsgroup.com/?l=samba-vms&r=1&w=2  > J > I found that one, also. Also found gname.org which provides a newsreader > interface to the list.   That's gmane.org. Sorry.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Mon, 26 May 2003 16:46:35 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton)& Subject: Re: Samba for OpenVMS How-To?. Message-ID: <LxrAa.730568$OV.673005@rwcrnsc54>  Z In article <3ED23F9C.F2BBC57@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes: <snip>G >I haven't gotten as far as SWAT yet. It kinda flies in the face of one C >of my secondary goals to make the kit as "VMS" as possible without F >requiring a lot of ancillary and/or non-VMS stuff to configure/manageC >it. I was thinking of a DCL proc. ala NET$CONFIGURE that could ask 0 >questions and re-write SMB.CONF as appropriate.  K Please forgive me; I don't mean to argue needlessly here.  I feel, however, N that as long as folks feel that GUI/web interfaces are somehow "anti-VMS", VMSL will remain "in the dark", hidden away from those who know nothing about it.  O You are one of the *many* advocates for wider use/acceptance of VMS; one of the G ways to increase use/acceptability of VMS is to give folks a "familiar" O interface to deal with.  Of course, there should still be CLI-level "tools" for O use; unfortunately, even HP has begun to move away from such dual-uses (witness - the lack of CLI for the HSV/SAN environment).   B There is/are CLI-level tools for Samba configuration, but they areO Unix-oriented (as are many networking-type products on VMS).  There seems to be  no escaping it.	:-)    >  >--  >David J. Dachtera >dba DJE Systems >http://www.djesys.com/  > ) >Unofficial Affordable OpenVMS Home Page:   >http://www.djesys.com/vms/soho/  A _________________________________________________________________ 0 Bradford J. Hamilton			"All opinions are my own"/ bMradAhamiPltSon@atMtAbi.cPoSm		"Lose the MAPS"    ------------------------------  % Date: Mon, 26 May 2003 11:45:29 -0500s1 From: "David J. Dachtera" <djesys.nospam@fsi.net> - Subject: SLS + DCSC + ACSLS: Spanning volumes ' Message-ID: <3ED244A9.119BC44C@fsi.net>w  6 Just wondering if anyone else out there has seen this.  D If you saw the earlier thread, you know that I've been locking hornsF with SLS, DCSC and ACSLS at work. One of the major challenges has been> the inexperience of the vendors with OpenVMS and fibre-channelF interconnects where everything after the HBA is non-HP kit. The switchE is a Brocade Sidewinder 12000, the NSRs are StorageTek SN3400 and then drives are Quantum SDLT-320s.   G Another major challenge has been getting SLS and DCSC to work correctlybC with ACSLS when backup savesets must span multiple volumes. It just  doesn't work right.   E From looking at the DCSC$TRACE_LOG, it appears that SLS/DCSC requests$C the eject of the current volume, but ACSLS responds some 26 secondsI; later with a status of 85 in the response to that request.    E It's hard to tell from the logs the exact sequence of events prior toI@ that. Based on past experience, it would seem that BACKUP does aH rewind/unload of the tape drive at EOV. Whether SLS waits until the tapeH reaches BOT and unloads before issuing the eject request to ACSLS is notD obvious from the logs. However, if you go by the time stamps, BACKUPG hits EOV some eight(8) minutes before DCSC issues the "dismount" (read:bH eject) request to ACSLS. The ACSLS error log indicates that the drive isD still busy. It shouldn't take eight(8) minutes to rewind a SDLT tape even from physical EOT.>  F So, dunno what's going on. Just wondering if anyone else has seen thisH in any other SLS+DCSC+ACSLS configuration and, if you did, did you solve; this problem and, if you did, what did the solution entail?l   -- m David J. Dachtera  dba DJE Systemse http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/i   ------------------------------  % Date: Mon, 26 May 2003 02:24:40 -0400s+ From: "Ron Milen" <milenronald@mailbag.net> ) Subject: Re: SYS$QIOW - Performance Issue # Message-ID: <3ed1b483$1_1@newsfeed>    Hi,uL     SYS$QIOW should work.  If the message is already in the mailbox then theL call should complete within milliseconds.  The only thing that comes to mindH is that there isn't any message waiting in the mailbox so the QIOW waitsI until it gets a message.  The documentation for the mailbox qio read call" can be found atdL http://h71000.www7.hp.com/doc/72final/6136/6136pro_013.html#mailbox_chap7 in chapter 4.3.1.  
 Good luck.	 Ron Mileno milenronald@yahoo.com3  ) "Vivek" <visoni@bmc.com> wrote in messageo) news:vd3arir9qlj671@corp.supernews.com...  > Hi,f > ? > I am reading data from the mailbox using the call SYS$QIOW( )  > ( > SYS$QIOW reads 1024 bytes in one call. >h; > Mail Box has max buffer of 2048 and message size as 1024., >dL > I am encountering a problem of performance in reading from mail box on VAXD > machine, it is also happening on some ALPHA boxes and not on some. > F > The time taken by this call is 4 minutes to 10 minutes for eac call. >?J > My Query is, if SYS$QIOW is the right call to read message from MAILBOX? >r4 > Can I use alternative call to improve performance. >o > Thanks > Vivekd >  >  >l >n >a   ------------------------------  % Date: Mon, 26 May 2003 03:37:28 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issueo( Message-ID: <3ED1C40C.19C03D4@istop.com>   Vivek wrote:( > SYS$QIOW reads 1024 bytes in one call. > ; > Mail Box has max buffer of 2048 and message size as 1024.t  F > The time taken by this call is 4 minutes to 10 minutes for eac call.  4 During those minutes, is the process in LEF state ?     K If you do SHOW DEV MBAxxx:/full, what is the "Reference Count" ? That shows K how many channels are assigned to it. If you have 2 processes competing for." read operations, this may explain.  N With mailboxes and QIO, a message written to the mailbox gets delivered to the4 channel who is first in the queue to read a message.  M So if you have 2 processes with READ pendings, and a appliaction that expectsoM one "command" message followed by one "data" message, then process 1 will getoM the command message, and process 2 will get the "data message" (when it wouldsK expect a command message), and process 1 then waits for a data message thatm, should have accompanied the command message.    N With QIOW, your process goes into LEF (waiting for event flag to be set) until the IO completes.gI If your process goes into some other state, (especially one starting with 0 RW___) it means that there is a serious problem.  N Does the process which writes to the maibox work normally, or does it also get stuck for that amount of time ?     K When the QIO does complete, does the status in the IO status block indicaten* success ? Is the data delivered properly ?  J Even on a all mighty microvax II, mailbox IOs are very fast and certaintly don't take minutes to complete.o   ------------------------------  % Date: Mon, 26 May 2003 14:35:15 +0530s From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issueb/ Message-ID: <vd3m06mv13j4b2@corp.supernews.com>e  
 Hi Ron/Mezie,r  I As far as I know there is only one process reading the mailbox, I need to  check again.  L If some process is writing to the mail box and other wants to read it, can I see it in reference count?.   K Most probably I will have to check if the block is written into the mailboxa or not as pointed by Ron.a   You have been a great help.    Thanks VivekY    7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messagea" news:3ED1C40C.19C03D4@istop.com... > Vivek wrote:* > > SYS$QIOW reads 1024 bytes in one call. > >t= > > Mail Box has max buffer of 2048 and message size as 1024.o >nH > > The time taken by this call is 4 minutes to 10 minutes for eac call. >M5 > During those minutes, is the process in LEF state ?v >i >tG > If you do SHOW DEV MBAxxx:/full, what is the "Reference Count" ? Thata showsyI > how many channels are assigned to it. If you have 2 processes competing. for'$ > read operations, this may explain. >nL > With mailboxes and QIO, a message written to the mailbox gets delivered to theo6 > channel who is first in the queue to read a message. >kG > So if you have 2 processes with READ pendings, and a appliaction thats expectsoK > one "command" message followed by one "data" message, then process 1 will  getcI > the command message, and process 2 will get the "data message" (when it  wouldnH > expect a command message), and process 1 then waits for a data message that. > should have accompanied the command message. >. > J > With QIOW, your process goes into LEF (waiting for event flag to be set) until- > the IO completes. K > If your process goes into some other state, (especially one starting withn2 > RW___) it means that there is a serious problem. >nL > Does the process which writes to the maibox work normally, or does it also get.! > stuck for that amount of time ?K >  >oD > When the QIO does complete, does the status in the IO status block indicate, > success ? Is the data delivered properly ? >sL > Even on a all mighty microvax II, mailbox IOs are very fast and certaintly! > don't take minutes to complete.v   ------------------------------  % Date: Mon, 26 May 2003 05:36:50 -0400n* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: SYS$QIOW - Performance Issuew) Message-ID: <3ED1E02F.8E68E778@istop.com>r   Vivek wrote:N > If some process is writing to the mail box and other wants to read it, can I > see it in reference count?.:  + Yes, you would see a reference count of 2. g  M But you could also have a single process generating a reference count of 2 ifLG it has 2 channels opened to that mailbox (for instance a lonely process K talking to itself with thr writes done on channel 1 and reads on channel 2.e  M > Most probably I will have to check if the block is written into the mailboxv > or not as pointed by Ron.>  0 Another thing you can do, from a SPARE terminal:   $OPEN /READ/WRITE TEMP MBAnnn:% $WRITE TEMP "Deadly chocolate mousse"i $CLOSE TEMPs  . See if this triggers the READ in your process.  L Note: if the maibox is full because the reads do not work, the above may getM stuck in the WRITE statement and you might need to kill the process. (or waitoL for the reading process to read from mailbox). But if that happens, it gives you an idea.  J Are you sure that the reading process has actually issued the QIOW ? Is itM possible it is getting stuck waiting for something else before getting to thenK QIOW ? (and when it does get to it, it comletes immetiatly because there iss already data in the mailbox)..   Another thing you can do:g   $OPEN/READ/WRITE TEMP MBAnnn:s $READ TEMP test  $SHOW SYMBOL TESTn $CLOSE TEMPA  H if the read completes right away, while your other main process is stillL taking 10 minutes, it is an indication that the other process had not queuedK the request to read before you queued yours, so you got the first packet inb the mailbox.  N If the read gets stuck, you can either kill the process or get another processF to write to the mailbox (see the example above) and that will free it.  6 This can help you *MAYBE* understand what is going on.  M Another thing to look at is the transaction count in the SHOW DEV/FULL to sees if it moves.   ------------------------------  % Date: Mon, 26 May 2003 15:35:00 +0530e From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issue2/ Message-ID: <vd3pg5c7hfdqab@corp.supernews.com>L  	 Hi Mezei,n   "Reference Count" is 2....   Thanks Viveko  7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messageg" news:3ED1C40C.19C03D4@istop.com... > Vivek wrote:* > > SYS$QIOW reads 1024 bytes in one call. > >a= > > Mail Box has max buffer of 2048 and message size as 1024.y > H > > The time taken by this call is 4 minutes to 10 minutes for eac call. >a5 > During those minutes, is the process in LEF state ?h >e >nG > If you do SHOW DEV MBAxxx:/full, what is the "Reference Count" ? Thatn showseI > how many channels are assigned to it. If you have 2 processes competingd fors$ > read operations, this may explain. > L > With mailboxes and QIO, a message written to the mailbox gets delivered to theo6 > channel who is first in the queue to read a message. >SG > So if you have 2 processes with READ pendings, and a appliaction thata expectsaK > one "command" message followed by one "data" message, then process 1 will  getcI > the command message, and process 2 will get the "data message" (when it2 would0H > expect a command message), and process 1 then waits for a data message that. > should have accompanied the command message. >4 >nJ > With QIOW, your process goes into LEF (waiting for event flag to be set) untila > the IO completes.oK > If your process goes into some other state, (especially one starting withe2 > RW___) it means that there is a serious problem. >0L > Does the process which writes to the maibox work normally, or does it also get>! > stuck for that amount of time ?a >  >aD > When the QIO does complete, does the status in the IO status block indicate, > success ? Is the data delivered properly ? >>L > Even on a all mighty microvax II, mailbox IOs are very fast and certaintly! > don't take minutes to complete.n   ------------------------------  % Date: Mon, 26 May 2003 11:08:02 +0100i" From: "Gary" <gmcd@totalise.co.uk>) Subject: Re: SYS$QIOW - Performance Issued. Message-ID: <basoru$s47$1@news8.svr.pol.co.uk>  L Have you used the io$m_now fuction code modifier for asychronuous read/write completion?     ) "Vivek" <visoni@bmc.com> wrote in messagep) news:vd3arir9qlj671@corp.supernews.com...  > Hi,( >T? > I am reading data from the mailbox using the call SYS$QIOW( )s >k( > SYS$QIOW reads 1024 bytes in one call. >-; > Mail Box has max buffer of 2048 and message size as 1024." >@L > I am encountering a problem of performance in reading from mail box on VAXD > machine, it is also happening on some ALPHA boxes and not on some. >.F > The time taken by this call is 4 minutes to 10 minutes for eac call. >.J > My Query is, if SYS$QIOW is the right call to read message from MAILBOX? >i4 > Can I use alternative call to improve performance. >e > Thanks > Vivekf >y >  >e >h >    ------------------------------  % Date: Mon, 26 May 2003 15:46:10 +0530  From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issuee/ Message-ID: <vd3q596i17phd9@corp.supernews.com>    This is what I did   $OPEN /READ/WRITE TEMP MBAnnn:% $WRITE TEMP "Deadly chocolate mousse"e $CLOSE TEMP    This is what it shows..J $OPEN /READ/WRITE TEMP MBA1    $READ TEMP test    $SHOW SYMBOL test    TEST = "Deadly chocolate mouse"0   $ CLOSE TEMP      7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messagea" news:3ED1C40C.19C03D4@istop.com... > Vivek wrote:* > > SYS$QIOW reads 1024 bytes in one call. > >e= > > Mail Box has max buffer of 2048 and message size as 1024.e >yH > > The time taken by this call is 4 minutes to 10 minutes for eac call. >-5 > During those minutes, is the process in LEF state ?P >  >rG > If you do SHOW DEV MBAxxx:/full, what is the "Reference Count" ? Thate showseI > how many channels are assigned to it. If you have 2 processes competingi for $ > read operations, this may explain. > L > With mailboxes and QIO, a message written to the mailbox gets delivered to the 6 > channel who is first in the queue to read a message. >nG > So if you have 2 processes with READ pendings, and a appliaction that  expects K > one "command" message followed by one "data" message, then process 1 willo getwI > the command message, and process 2 will get the "data message" (when ito wouldeH > expect a command message), and process 1 then waits for a data message that. > should have accompanied the command message. >l >nJ > With QIOW, your process goes into LEF (waiting for event flag to be set) untila > the IO completes.lK > If your process goes into some other state, (especially one starting withs2 > RW___) it means that there is a serious problem. >oL > Does the process which writes to the maibox work normally, or does it also gety! > stuck for that amount of time ?r >  >aD > When the QIO does complete, does the status in the IO status block indicate, > success ? Is the data delivered properly ? >iL > Even on a all mighty microvax II, mailbox IOs are very fast and certaintly! > don't take minutes to complete.-   ------------------------------  % Date: Mon, 26 May 2003 20:40:08 +0530. From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issue / Message-ID: <vd4bce3i58apc3@corp.supernews.com>0  	 Hi  Gary,m  < All the SYS$QIOW calls only had the WRITEVBLK function code.  G After adding the WRITEBVBLK function code the time take is reduced to 1i minutes from 4-5 minutes.e   Seems there is and improvement.?   Thanks Vivekr      - "Gary" <gmcd@totalise.co.uk> wrote in messageh( news:basoru$s47$1@news8.svr.pol.co.uk...C > Have you used the io$m_now fuction code modifier for asychronuousa
 read/write
 > completion?d >D >v+ > "Vivek" <visoni@bmc.com> wrote in messagee+ > news:vd3arir9qlj671@corp.supernews.com...( > > Hi,/ > > A > > I am reading data from the mailbox using the call SYS$QIOW( )  > >w* > > SYS$QIOW reads 1024 bytes in one call. > >e= > > Mail Box has max buffer of 2048 and message size as 1024.h > >pJ > > I am encountering a problem of performance in reading from mail box on VAX F > > machine, it is also happening on some ALPHA boxes and not on some. > > H > > The time taken by this call is 4 minutes to 10 minutes for eac call. > >-L > > My Query is, if SYS$QIOW is the right call to read message from MAILBOX? > >m6 > > Can I use alternative call to improve performance. > >D
 > > Thanks	 > > Vivek7 > >c > >t > >o > >c > >  >u >s   ------------------------------  % Date: Mon, 26 May 2003 20:25:03 +0530r From: "Vivek" <visoni@bmc.com>) Subject: Re: SYS$QIOW - Performance Issuee/ Message-ID: <vd4bccrr8pemc0@corp.supernews.com>u  	 Hi  Gary,e  3 I hav now added the IO$M_NOW function code modifer.l   Thanks Viveks      - "Gary" <gmcd@totalise.co.uk> wrote in message ( news:basoru$s47$1@news8.svr.pol.co.uk...C > Have you used the io$m_now fuction code modifier for asychronuoust
 read/write
 > completion?p >  > + > "Vivek" <visoni@bmc.com> wrote in messageh+ > news:vd3arir9qlj671@corp.supernews.com...o > > Hi,o > >kA > > I am reading data from the mailbox using the call SYS$QIOW( )f > >r* > > SYS$QIOW reads 1024 bytes in one call. > > = > > Mail Box has max buffer of 2048 and message size as 1024.. > > J > > I am encountering a problem of performance in reading from mail box on VAXeF > > machine, it is also happening on some ALPHA boxes and not on some. > >cH > > The time taken by this call is 4 minutes to 10 minutes for eac call. > > L > > My Query is, if SYS$QIOW is the right call to read message from MAILBOX? > > 6 > > Can I use alternative call to improve performance. > > 
 > > Thanks	 > > Vivek  > >d > >4 > >o > >s > >o >3 >f   ------------------------------  % Date: Mon, 26 May 2003 11:31:22 -0400n' From: Howard S Shubs <howard@shubs.net>p) Subject: Re: SYS$QIOW - Performance Issuer< Message-ID: <howard-221203.11312226052003@enews.newsguy.com>  . In article <basoru$s47$1@news8.svr.pol.co.uk>,$  "Gary" <gmcd@totalise.co.uk> wrote:  N > Have you used the io$m_now fuction code modifier for asychronuous read/write
 > completion?l  H Uh, w/o looking at the docs first, why would he do that instead of just  using SYS$QIO w/o the W?   -- e4 Today, on Paper-view: The World Origami Championship   ------------------------------  % Date: Mon, 26 May 2003 10:10:06 -0400u. From: Dave Bossi <davebossi@C.O.M.C.A.S.T.net>; Subject: Re: ZLE data synchronise with different line speeda0 Message-ID: <3ED2203E.2050902@C.O.M.C.A.S.T.net>  H Actually this issue has nothing to do with ZLE, it could occur any time B two clients try to update the same data.  This is why application D programmers either code in locking mechanisms or implement logic to . validate the original image prior to updating.  D Do not confuse a ZLE environment with permitting more than a single F "database of record" for each application.  There are cases where the H ODS may be that database, or more commonly in the "real" world, the ODS G is simply the collection point for whole or partial data replicated in  1 near real time from existing legacy applications.t   Dave   MAser wrote:H > Thank for the reply but I seem to have misguide you about my question. > F > I was not thinking on the core side of ZLE but on the client side orH > the application. Like I want to access data and update it from my homeH > (v.90)and at the same time someone else access from a T1 line the sameH > data and do some updating. I think that due to the difference in speedF > and access data time. I will not able to get the latest data or even > updating it to the latest. >  >  >  > "Randall S. Becker" <r s b e c k e r _ n o s p a m@n e x b r i d g e _n o s p a m . c o m> wrote in message news:<F5Yza.254317$w7k.214182@news04.bloor.is.net.cable.rogers.com>... > L >>Another thing to consider is that a V.90 modem isn't going to be up to theN >>task of handling a 1000 or 10000 TPS load unless the transactions are fairlyH >>trivial - in which case, why bother with a ZLE in the first place? TheM >>infrastructure required to support a true ZLE is substantial, and is beyondf! >>simple data transmission rates.s >>= >>"Dave Bossi" <davebossi@C.O.M.C.A.S.T.net> wrote in messagee, >>news:3ECE1F11.3080701@C.O.M.C.A.S.T.net...G >>| The Zero Latency concept applies to a much greater picture than anyeI >>| individual piece of the replication activity.  If you go to HP's siterL >>| you can read information on ZLE and what the concept really encompasses. >>|0J >>| Regarding replication over a slow circuit, you are correct, the higherI >>| speed the circuit the more current the data that has been replicated.dL >>| Since replication tools work "in the background", your real-time updatesK >>| to the primary site are not affected by the bandwidth to the remote. TotK >>| establish a mirror site for DR or for load sharing, your bandwidth must L >>| be at least equal to your aggregate data modification rate or the mirror >>| site will fall behind. >>|c >>| Dave >>|s >>| MAser wrote:E >>| > I was wonder about this ZLE concept from hp. It say that it canTK >>| > synchronise data. I was thinking if the speed of the line does affectoF >>| > the data input. For example, if I were to use normal phone modemK >>| > (v.90) and access & update the database. At the same time, if anotherrJ >>| > person used DSL and access the same data would there be a problem as. >>| > now mostly, he will read the older data. >>| >  >>| > Thank youh >>|y >    ------------------------------   End of INFO-VAX 2003.290 ************************