1 INFO-VAX	Sat, 20 May 2006	Volume 2006 : Issue 279       Contents:" Re: Fixing a Corrupt PCSI Database Re: FreeVMS New release 0.2.10J Re: How to restore one BACKUP/IMAGE from tape with mutliple BACKUP/IMAGEs?P Re: How to restore one BACKUP/IMAGE from tape with mutliple BACKUP/IMAGEs? BACKU+ Re: Multiple job-openings in Sacramento, CA P OT: Woodcrest (X86-64)  will ouperform all other cpus on the market says InquireP Re: OT: Woodcrest (X86-64) will ouperform all other cpus on the market says Inqu Re: Performance and Disk Size # Re: Ping in web server command file # Re: Ping in web server command file  Re: VPM_SERVER strangeness/ Re: what should Bad-Clients: in SMTP.CONFIG do?   F ----------------------------------------------------------------------  % Date: Sat, 20 May 2006 12:00:14 -0400 + From: Steve Matzura <number6@speakeasy.net> + Subject: Re: Fixing a Corrupt PCSI Database 8 Message-ID: <q5fu62tcd2bmobnccd5atltapa6s96e83m@4ax.com>  . On Fri, 19 May 2006 18:27:11 GMT, Hoff Hoffman  <hoff-remove-this@hp.com> wrote:  F >   I wish you luck with your recommendation, but what those HP folks E >told you was very likely correct, and likely the safest course.   I  B >would tend to expect to have to perform a roll forward to regain A >consistency, an approach the folks you talked with at HP likely   >commented on.  D No, they did not.  I'll ask about it.  Any info before I get to them would be greatly appreciated.    ------------------------------    Date: 20 May 2006 05:40:05 -0700! From: "mariuz" <mapopa@gmail.com> ' Subject: Re: FreeVMS New release 0.2.10 B Message-ID: <1148128805.630979.70840@j55g2000cwa.googlegroups.com>  $ Another alternative to bochs is qemu$ Download the images from above links   gunzip a.img.gz  gunzip c.img.gz   1 qemu -fda a.img -hda c.img -boot a -monitor stdio    Select the default grub option  . You can read mor in the USE link (on the site)   ------------------------------  % Date: Sat, 20 May 2006 08:08:51 +0200 3 From: Wilm Boerhout <w4OLD.boerhout@PAINTplanet.nl> S Subject: Re: How to restore one BACKUP/IMAGE from tape with mutliple BACKUP/IMAGEs? 4 Message-ID: <446eb273$0$988$ba620dc5@nova.planet.nl>  A and, if you're not sure about the names, you can always list the   contents of the tape by    $MOU /OVER=ID MKcuu: $DIR /COL=1 MKcuu:   /Wilm    ------------------------------  % Date: Fri, 19 May 2006 22:36:26 -0700 ( From: Jeff Cameron <roktsci@comcast.net>Y Subject: Re: How to restore one BACKUP/IMAGE from tape with mutliple BACKUP/IMAGEs? BACKU 0 Message-ID: <C093F8EA.1F7DF%roktsci@comcast.net>  J On 5/19/06 9:31 PM, in article vYwbg.186$gc6.110@fe06.lga, "Z" <Z@ids.net> wrote:  H > I have a backup tape. It contains 10 $BACKUP/IMAGEs, written one after3 > the other, one for each disk on the Alpha server.  > J > I need to restore just one disk from the tape, the 6th disk backed up on > the tape.  >  > What are the proper commands?  > M > Do I need to skip over savesets/images? Or just pick the right one by name?  > . > This is VMS 6.2, if that makes a difference.  J  In the example below the target disk is $1$DKA100:, and the tape drive is: $4MKB200:, and the image backup saveset is named DATA.BCK.  < 1. Mount the target disk device with the /FOREIGN qualifier:    $MOUNT/FOREIGN $1$DKA100:  . 2. Mount the tape with the /FOREIGN Qualifier.    $MOUNT/FOREIGN $4$MKB200:   3. Do the restore 3    $BACKUP/IMAGE $4$MKB200:DATA.BCK/SAVE $1$DKA100:   F This will automatically start searching the tape for the saveset namedD DATA.BCK. You do not have to position the tape. If you have multipleJ savesets named the same then you will have to position the tape before you do the backup command.&    $SET MAGTAPE/SKIP=FILE=n $4$MKB200:  C Calculate the "n" using the formula:  n = (Save_set_number * 3) - 2    Jeff   ------------------------------    Date: 20 May 2006 07:55:18 -0700 From: bob@instantwhip.com 4 Subject: Re: Multiple job-openings in Sacramento, CAB Message-ID: <1148136918.557382.287450@38g2000cwa.googlegroups.com>  ' going to go from no viruses or hacks to ( the patch of the week club ... brilliant( thinking ... just watch the TCO rise ...   ------------------------------  # Date: Sat, 20 May 2006 13:31:22 GMT ( From: Alan Greig <greigaln@netscape.net>Y Subject: OT: Woodcrest (X86-64)  will ouperform all other cpus on the market says Inquire > Message-ID: <KSEbg.304284$8Q3.15423@fe1.news.blueyonder.co.uk>  6 At least on floating point Linpack benchmark anyway...  ) http://www.theinquirer.net/?article=31836 / By Nebojsa Novakovic: Friday 19 May 2006, 21:53   I A FEW DAYS AGO, a friend told me that the famous Linpack FLOPs benchmark  G for matrix multiplication, still used (despite its near uselessness in  G real apps) to rank the world's TOP 500 supercomputers, will run on the  A Conroe and Woodcrest far better than expected, due to the recent   compiler and tuning advances.   D Basically, a 3GHz Woodcrest chip gives you 3 billion x 4 FP ops x 2 A cores per second, or 24 GFLOPS theoretical peak (Rpeak number in  ( Linpack) per socket in 64-bit precision.  C So, two chips on a typical workstation or server board give you 48  F GFLOPS Rpeak - almost as good as a quad-chip Montecito, or 50% better * than dual-chip (four cores total) POWER5+.  G With the rumoured improvements, Woodcrest is hitting 80% efficiency in  H Linpack, i.e. its measurable Rmax rate of execution will be four-fifths 4 of theoretical peak, or over 38 GFLOPs in this case.  I Now, the 2.8GHz dual-core Opteron grade, expected to be the one greeting  G Woodcrest, has some 88% efficiency per socket, but the Rpeak per clock  H is half. So, it is 2.8 billion x 2 FP ops x 2 cores per second, or 11.2 I GFLOPs Rpeak per socket. For two-socket Opteron, the Rpeak then would be  D 22.4 GFLOPs, and Rmax some 19.8 GFLOPs - half that of Woodcrest two E socket setup! Wow, AMD should have had K8L now, not in a year's time.   I We all know that Itanic is darn great at least in Linpack, with over 90%  @ efficiency in some cases - but its clock is nearly half that of B Woodcrest, with the same peak FP rate per cycle anyway. So, for a I dual-core 1.6GHz Montecito, we have 1.6 billion x 4 FP ops x 2 cores per  E second, or 12.8 GFLOPs Rpeak per socket. Times 0.9 for Rmax, we have  L 11.5 GFLOPs Rmax per socket, or roughly 23 GFLOPs Rmax per two-socket board.  E And 2.2GHz Power5+? Let's say 77% efficiency in this case. Dual-core  I chip too, so 2.2 billion x 4 FP ops x 2 cores per second, or 17.6 GFLOPs  J Rpeak, and 13.5 GFLOPs Rmax per socket, or 27 GFLOPs per two-socket board.  I In summary? Woodcrest will, socket for socket, outperform all other CPUs  H in the market, X86 or RISC or EPIC - when it comes to the TOP500 battle  and its Linpack benchmark.  @ We all also know that the real application performance is quite I different from Linpack benchmark code, and that the good ship Itanic, as  E well as POWER4, have more registers to play with, and more efficient  G instruction set architectures than X86-64, so there will be apps where  I either of these two will still outperform Woodcrest or Opteron. And yes,  I the new Core 2 Duo doesn't have the quad-socket MP option yet for larger   compute nodes.  G But, with almost deadly assurance, I can state that the number of such  H apps will be greatly diminished - at least in the workstation and small C server node class of supercomputing apps. You don't really need to  6 re-optimised your apps to enjoy most of the new power.  G Importantly, most tender bids for large supercomputer clusters include  E the Linpack number TOP500 position in a very prominent place. It's a   question of bragging rights.    --  
 Alan Greig   ------------------------------    Date: 20 May 2006 07:52:00 -0700 From: bob@instantwhip.com Y Subject: Re: OT: Woodcrest (X86-64) will ouperform all other cpus on the market says Inqu C Message-ID: <1148136720.617884.260990@y43g2000cwc.googlegroups.com>   # so when does the vms port start? :)    ------------------------------  % Date: Sat, 20 May 2006 20:54:25 +0800  From: prep@prep.synonet.com & Subject: Re: Performance and Disk Size0 Message-ID: <87y7ww26wu.fsf@k9.prep.synonet.com>  . Hoff Hoffman <hoff-remove-this@hp.com> writes:   > Dan Moore wrote:  @ >>    We are moving from 18GB (Ultra 320 15K) disks to something> >> larger (36, 72, or 144). Assuming nothing else changes, I'm< >> interested in the performance impact that could result by >> increasing the disk size.  G It will depend to a large degree on how you arange your data on the new 4 disks, and on what your current access patterns are.  F If, for example, you put all the data from 4 18s onto a single 144 youD will now have all the IOs queued to s single head. This can suck bigD time, and can really throw you when you do the change over. Moving xE drives of 18 GB to x 144s will almost always be a win. Same number of C heads, for no extra contention there, and a lot more data under the D heads with out having to do a seak. And the seaks will on average be shorter.  B If performance is your need, keeping the number of drives to a bus7 down to 2 or 3 is still the way to go as far as I know.   F >    A few of the higher-end I/O controllers around can and do benefitA > from larger and from larger, and naturally-aligned disk cluster F > factors. (OpenVMS ODS-2 and ODS-5 always have naturally-aligned diskC > clusters.) There are also changes planned for and arriving within A > newer disks, particularly around the native (internal) blocking F > within the drives; the internal block sizes are increasing, and thusB > the disks that have these larger internal sectors can themselvesF > perform better with aligned transfers.  Accordingly and as a genericC > recommendation, I tend to use and to encourage 16, 32 or 64 block B > cluster factors when initializing a disk.  (This one is from the > controller I/O gurus.)  D Is this a recent change? I inited an 18 GB drive 2 days ago, then reA did it to get rid of the 35 block CF and did it at, ah 16 or 32 I E think. Power of 2 anyway. Extention is 1024, photos tend to have that C effect on you :), and default IOs are 64. All of this done by hand.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 20 May 2006 05:14:38 -0700( From: "Scott" <DocTrinsograce@gmail.com>, Subject: Re: Ping in web server command fileC Message-ID: <1148127278.255768.133030@u72g2000cwu.googlegroups.com>    Bob Koehler wrote:F >    Technically it's all binary.  But how would you say non-HTML text, >    without the clumsy phrase I just wrote?  C Sometimes being explicit is inherently clumsy.  Its what we do. :-) 3 Does "simple text" trip more lightly over the gums?    ------------------------------  % Date: Sat, 20 May 2006 08:47:09 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca> , Subject: Re: Ping in web server command file: Message-ID: <EdEbg.10512$aa4.414154@news20.bellglobal.com>  I "Bob Koehler" <koehler@eisner.nospam.encompasserve.org> wrote in message  - news:iGk9D4mmF$ZL@eisner.encompasserve.org... I > In article <Fk4bg.39092$Hk1.27385@read1.cgocable.net>, "Mike Robinson"   > <onthost@gmail.com> writes:    [...snip...]   > E >   Technically it's all binary.  But how would you say non-HTML text + >   without the clumsy phrase I just wrote?  >   K How about "PLAIN TEXT"? (since this is a MIME type that the HTML poster is   probably familiar with)   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada." http://www3.sympatico.ca/n.rieck/    ------------------------------  + Date: Sat, 20 May 2006 12:56:44 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)# Subject: Re: VPM_SERVER strangeness $ Message-ID: <e4n3mc$o5a$1@online.de>  = In article <Crkbg.808$7u2.460@news.cpqcorp.net>, Hoff Hoffman " <hoff-remove-this@hp.com> writes:   1 > Phillip Helbig---remove CLOTHES to reply wrote: + > > ...VPM server was showing a few hundred + > > thousand page faults and 0 CPU usage... H > > Another cluster, same VMS and patch level, shows no problems at all.K > > I restarted the VPM server on the system with problems.  No change: it  , > > immediately starts throwing page faults. > J >    That two OpenVMS systems are at the same revision level is certainly H > interesting and useful, but knowing that both are at the most current H > level is arguably more interesting.  (The former implies a difference J > between the box that doesn't involve ECO kits, while the latter implies H > the potential for a new and as-yet unresolved problem within OpenVMS.)  E They are both 7.3-2, fully patched (except for the latest FIBRE_SCSI   patch).    >    VPM via DECnet or IP?   IP.    >    OpenVMS version     See above.     > and architecture?   E The system with the problem is COMPAQ Professional Workstation XP1000 @ while the one in the same cluster without the problem is DigitalG AlphaStation 500/500.  (In my cluster, I have a DEC 3000/600 and a 5305 C (ALPHAserver 1200)---both at 7.3-2 and fully patched (the 5305 is a < satellite of the DEC 3000)---and have not seen the problem.)   >    How was VPM started?   + Via the startup database feature in SYSMAN.   I >    Different system default quotas?  (That case shouldn't hit VPM, but  J > I've seen that trip various applications over the years -- depending on D > the system default quotas can lead to odd application differences 9 > between two otherwise identical systems, for instance.)    There is a common SYSUAF.   D >    There are ECOs available for VPM-related problems for specific  > OpenVMS releases.   G The problem is still there.  A restart of the VPM_SERVER didn't change  D anything.  Here are some outputs from SHOW SYSTEM taken a couple of + seconds apart.  Note the lack of CPU usage.   M 2120014E VPM_SERVER      HIB     15       18   0 00:00:00.02   9201890    107   M 2120014E VPM_SERVER      HIB     15       18   0 00:00:00.02   9203060    107   M 2120014E VPM_SERVER      HIB     15       18   0 00:00:00.02   9204464    107   M 2120014E VPM_SERVER      HIB     15       18   0 00:00:00.02   9205400    107   G SHOW PROC/CONT shows only the page faults changing, everything else is   static.   G (This isn't my system but I do have access to a priviledged account on  ' it, so could try out some diagnostics.)   ) SHOW ERROR doesn't show anything of note.   L Physical Memory Usage (pages):     Total        Free      In Use    ModifiedL   Main Memory (256.00MB)           32768       21441        9526        1801   ------------------------------  + Date: Sat, 20 May 2006 13:02:07 +0000 (UTC) P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)8 Subject: Re: what should Bad-Clients: in SMTP.CONFIG do?$ Message-ID: <e4n40f$o5a$2@online.de>  6 In article <446E2A57.39652BB3@vaxination.ca>, JF Mezei  <jfmezei@vaxination.ca> writes:   H > If I connect to your system, it is because I want to deliver a spam to: > one of your users. OR, use your SMTP server as a relay.    Right.  I > If I want to deliver to one of your users, then I only connect you your P > system when the email address has a host whose MX entry points to your system.   Right.  W > > > Client IP address 211.213.103.30 unbacktranslatable (gethostbyaddr returned NULL)  > 
 > Do you have & > 		Reject-Unbacktranslatable-IP: TRUEG > in your SMTP.CONFIG ? If not, you should. That will stop reception of  > the message at that point.  F No.  I can understand people who ALWAYS have spam-blocks in the From: I address, and I don't want to drop those.  Of course, the place to put it  F is in the domain, not in the username, so that the domain doesn't get  mail for a non-existent user.   8 > > This is the address I have in the Bad-Clients: list. > M > I am not sure which is tested first, unbacktranslatable or the bad-clients.   ) Mail from unbacktranslatable is accepted.   - > > Printing debug_level 2, relevant headers: ' > > Return-Path: SMTP%"dtree7@nate.com" < > > Received: from unknown.hostname (211.213.103.30) by xxxx > > 3 > > Printing debug_level 2, Domains and recipients:  > > Domain: nate.com0 > >    Recipient address: SMTP%"dtree7@nate.com"" > >    Domain part:       nate.com  > >    Local part:        dtree71 > >    Address Status:    Done, delivered. (Sent)  > G > What this tells me is that your SMTP server allows relaying from some F > unknown hostname to an address that is not within your jurisdiction. > VERY BAD !  F No, relaying is switched off.  This was an example of a mail from the I same IP address to a domain I DO accept mail for.  Since the user was a)  B non-existent and b) more than 12 characters long, the message was > bounced.  The stuff above shows the address it was bounced to.  F > So it appears that the receiver will allow the connection, issue theF > message and disconnect afterwards. This does generate a receiver logJ > file since the receiver is in fact started and reads the smtp config andH > applies the rules. So yes, a search in your log files will reveal that > many entries for blocked IPs.   ) Today or tomorrow, I will test in detail.   I > This page contains lots of links on relay testing. You can look at some F > of the web based relay checkers and have them test your SMTP server.8 > They will tell you all your configuration weakenesses.   Thanks.    ------------------------------   End of INFO-VAX 2006.279 ************************