1 INFO-VAX	Wed, 14 Dec 2005	Volume 2005 : Issue 694       Contents:" Re: calloc() fails, no more memory) Re: DS10L hardware configuration question  Re: ES40 CPU Upgrade6 Re: Extreme backup slowdown after ECO updates (V7.3-1)6 Re: Extreme backup slowdown after ECO updates (V7.3-1)6 Re: Extreme backup slowdown after ECO updates (V7.3-1) java OutOfMemoryError  Re: java OutOfMemoryError * Question about MiniMerge/MiniCopy Bitmaps.. Re: Question about MiniMerge/MiniCopy Bitmaps.A Re: Question about SPAWN/INPUT table in the DCL Dictionary manual % Re: SHOW DEVICE: what should it show? % Re: SHOW DEVICE: what should it show?  Re: SRM memtest 0 RE: Timeout strategy: terminal vs Telnet drivers0 Re: Timeout strategy: terminal vs Telnet drivers  F ----------------------------------------------------------------------  % Date: Tue, 13 Dec 2005 19:37:18 +0000 " From: Hans Blom <tocum2@yahoo.com>+ Subject: Re: calloc() fails, no more memory J Message-ID: <1134502638.3dab6d17bb59dd43e6b8995f6fc02cd0@fe2.teranews.com>   Hans Blom wrote: > Hello all,I > I'm running OpenVMS 7.3-2 on an Alphaserver. A programmer has developed I > an application that, in order to work with speed, wants to keep as much I > as possible of the data in memory. At some point he does an calloc() to F > get a memoryarea to keep about 400 000 pointers. He gets a null backF > from the call, basically OpenVMS saying - sorry sir, no more memory! > J > We have tried raising pgflquota, wsextent (in order to decrease need forI > paging) and every other conceivable quota both in sysuaf and sysgen. We I > can see that as long as the program works in memory everything is fine, F > but as soon as wsextent is hit and it has to start paging, it fails. > 1 > I'm stuck! Anybody got any ideas on what to do?  > 	 > Regards  >  > Hans Blom E Hello all I think my answer went too deep into the nested dialogue so 
 I'm reposting    Hi all, & it grew to quite a thread. Thanks all.E I tried a loop calloc:ing 400000 pointers (1.6 Mb approx). The 383:rd 8 attempt fails. errno = %x0c, vaxc$errno = 0. I think theC programmer/developer should do a review of his program,  the calloc ? failing can't be the main problem, and then he is welcome back.    Thanks again all	 Hans Blom    ------------------------------  % Date: Tue, 13 Dec 2005 15:20:16 -0500 C From: "David Turner, Island Computers US Corp" <dbturner@icusc.com> 2 Subject: Re: DS10L hardware configuration question8 Message-ID: <QWFnf.5892$kP5.2697@bignews5.bellsouth.net>  L Of course, we sell them and I can even boot VMS 7.2-1 and 7.3-2 from a flash card (SD and COMPACTFLASH)0 But they are kinda slow compared to a USB device   --     David B Turner Island Computers US Corp 2700 Gregory St, Suite 180 Savannah GA 31404  Tel: 912 447 6622 X201 Cell: 912 447 6622 X252  Fax: 912 201 0402  Email: dbturner@icusc.com  Web: http://www.islandco.com% ===================================== < All orders are subject to the following terms and conditions. of sale. These should be read before ordering.% http://www.islandco.com/warranty.html   I "Tim ffrench-Lynch" <tim-DOT-ffrench-HYPHEN-lynch@selex-sas.com> wrote in / message news:439F077D.9960160E@selex-sas.com... 1 > "David Turner, Island Computers US Corp" wrote:  > > J > > Wouldn't mind some kind of USB drivers that recognize flash devices as hard	 > > disks  > - > Would an IDE to Compact Flash adaptor work?  > + > http://www.mini-itx.com/store/?c=16#p2602  >  > Tim    ------------------------------  # Date: Wed, 14 Dec 2005 04:41:14 GMT  From: dittman@dittman.net  Subject: Re: ES40 CPU Upgrade , Message-ID: <KnNnf.45833$qF6.19094@trnddc01>  B David Turner, Island Computers US Corp <dbturner@icusc.com> wrote:K <dittman@dittman.net> wrote in message news:0aknf.37070$Y7.6921@trnddc02... E >> I have an ES40 with two 667MHz CPUs.  The CPU part number for both E >> CPUs is 54-30362-01.A03.  These are EV67 pass 2.3.3.  If I add two C >> more 667MHz CPUs do I need to match the CPU part number and EV67 C >> pass?  I know this was an issue with the 500MHz CPUs but I can't ) >> find any warnings for the 667MHz CPUs.  >>B >> This is a hobbyist system so asking HP support isn't an option.  " > Yes - they need to be identical?  . That makes things a bit more complicated.  :-( --   Eric Dittman dittman@dittman.net    ------------------------------    Date: 13 Dec 2005 10:43:33 -0800( From: "Rich Jordan" <jordan@ccs4vms.com>? Subject: Re: Extreme backup slowdown after ECO updates (V7.3-1) C Message-ID: <1134499412.951049.217530@g47g2000cwa.googlegroups.com>   F Found.  The SYS$PKWDRIVER included in the VMS731 FIBRE_SCSI V7 kit hasD a bug in it; this kit is the most recent release.  HP has a remedialD image available if you get the shortblock errors on verification (or read) of the tape.  D Although the bug is in a mandatory (class 2, but if you use SCSI...)? ECO, HP is not planning on an updated ECO because V7.3-1 is off  support.  @ Although their notes showed the problem only happened with largeD blocksizes (60K+) we were seeing the bug with 8K, 32K, and 64K block sizes.  A Since installation of the new driver we have one full, successful G backup to tape.  We're verifying that tape on another system as soon as + possible, but it looks like problem solved.    Rich CCS    ------------------------------  % Date: Tue, 13 Dec 2005 14:37:36 -0500  From: norm.raphael@metso.com? Subject: Re: Extreme backup slowdown after ECO updates (V7.3-1) Q Message-ID: <OF6E75BFF4.CE07E625-ON852570D6.006BBDF6-852570D6.006BD081@metso.com>   C "Rich Jordan" <jordan@ccs4vms.com> wrote on 12/13/2005 01:43:33 PM:   H > Found.  The SYS$PKWDRIVER included in the VMS731 FIBRE_SCSI V7 kit hasF > a bug in it; this kit is the most recent release.  HP has a remedialF > image available if you get the shortblock errors on verification (or > read) of the tape. >   B Do you have a Problem Reference so [we] can contact HP about this?    F > Although the bug is in a mandatory (class 2, but if you use SCSI...)A > ECO, HP is not planning on an updated ECO because V7.3-1 is off 
 > support. > B > Although their notes showed the problem only happened with largeF > blocksizes (60K+) we were seeing the bug with 8K, 32K, and 64K block > sizes. > C > Since installation of the new driver we have one full, successful I > backup to tape.  We're verifying that tape on another system as soon as - > possible, but it looks like problem solved.  >  > Rich > CCS  >    ------------------------------    Date: 13 Dec 2005 15:35:38 -0800( From: "Rich Jordan" <jordan@ccs4vms.com>? Subject: Re: Extreme backup slowdown after ECO updates (V7.3-1) C Message-ID: <1134516938.384623.144350@g49g2000cwa.googlegroups.com>    norm.raphael@metso.com wrote: E > "Rich Jordan" <jordan@ccs4vms.com> wrote on 12/13/2005 01:43:33 PM:  > J > > Found.  The SYS$PKWDRIVER included in the VMS731 FIBRE_SCSI V7 kit hasH > > a bug in it; this kit is the most recent release.  HP has a remedialH > > image available if you get the shortblock errors on verification (or > > read) of the tape. > >  > D > Do you have a Problem Reference so [we] can contact HP about this? >  >   > According to Volker Halle's post to me on the ITRC forums, theC reference number is QXCM1000235564.  That did not come up during my 
 service call.    ------------------------------    Date: 13 Dec 2005 13:14:05 -0800' From: "tadamsmar" <tadamsmar@yahoo.com>  Subject: java OutOfMemoryErrorC Message-ID: <1134508445.066639.236590@g14g2000cwa.googlegroups.com>   ( I am getting this on a VMS alpha system.  B If it is due to resources (not just a bug in my program) what do I, do to provide more resources to the program?   Thanks in advance.   ------------------------------  + Date: Tue, 13 Dec 2005 22:15:23 +0000 (UTC) - From: klewis@OMEGA.MITRE.ORG (Keith A. Lewis) " Subject: Re: java OutOfMemoryError, Message-ID: <dnnh5r$r$1@newslocal.mitre.org>   "tadamsmar" <tadamsmar@yahoo.com> writes in article <1134508445.066639.236590@g14g2000cwa.googlegroups.com> dated 13 Dec 2005 13:14:05 -0800: ) >I am getting this on a VMS alpha system.  > C >If it is due to resources (not just a bug in my program) what do I - >do to provide more resources to the program?   L What version of java?  I think 1.4.2 has a default maximum heap size of 64M.  B You can increase that value by adding it as a command line option: $ java "-Xmx200m" 7 (You need quotes around the argument to preserve case.)   I Do you think your program should be using 64 MB?  If not, it's probably a  bug on your end.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------    Date: 13 Dec 2005 11:40:21 -0800" From: dave.baxter@bannerhealth.com3 Subject: Question about MiniMerge/MiniCopy Bitmaps. B Message-ID: <1134502821.208040.51530@g49g2000cwa.googlegroups.com>  B I have MiniMerge enabled on ~50 shadowsets, on 3 nodes in an alpha cluster (Ovms7.3-2).C I want to dismount/policy=minicopy, one unit out of each shadowset.     B My question is, is there any way that MiniMerge can interfere with MiniCopy.?????  F My shadowsets only have 2 members, so logically, I am guessing that asF soon as I dismount one with the MiniCopy Option, the MiniMerge becomes moot.    Dave.    ------------------------------    Date: 13 Dec 2005 15:57:28 -0600. From: brooks@cuebid.zko.hp.nospam (Rob Brooks)7 Subject: Re: Question about MiniMerge/MiniCopy Bitmaps. , Message-ID: <tnU3MnKq6JA5@cuebid.zko.hp.com>  $ dave.baxter@bannerhealth.com writes:  D > I have MiniMerge enabled on ~50 shadowsets, on 3 nodes in an alpha > cluster (Ovms7.3-2).  E > I want to dismount/policy=minicopy, one unit out of each shadowset.  > D > My question is, is there any way that MiniMerge can interfere with > MiniCopy.?????  L I'm not sure what you mean by "interfere", but I am quite sure the answer is no.   H > My shadowsets only have 2 members, so logically, I am guessing that asH > soon as I dismount one with the MiniCopy Option, the MiniMerge becomes > moot.   J Yes, although the minimerge bitmaps will remain, and continue to be active (registering write).   --    L Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.hp.com   ------------------------------    Date: 13 Dec 2005 12:45:58 -0600 From: briggs@encompasserve.orgJ Subject: Re: Question about SPAWN/INPUT table in the DCL Dictionary manual3 Message-ID: <ekFHiBXt0YPh@eisner.encompasserve.org>   i In article <1134487795.605868.89250@g49g2000cwa.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes: ! > briggs@encompasserve.org wrote: m >> In article <1134423071.479544.144610@o13g2000cwo.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:  >> >$ >> > briggs@encompasserve.org wrote:p >> >> In article <1134403065.283312.274140@z14g2000cwz.googlegroups.com>, "AEF" <spamsink2001@yahoo.com> writes:K >> >> > Apparently the confusion is coming from the NRO term. I thought all L >> >> > these examples you show use record-oriented devices. So what's going? >> >> > on? I think this term, NRO, is the crux of the problem.  >> >> D >> >> Disk devices are not record oriented.  They are file oriented. >> >> O >> >> Disk EISNER$DRA3:, device type 7 Member RAID 5, is online, mounted, file- U >> >>     oriented device, shareable, available to cluster, error logging is enabled.  >> >> S >> >> Device MBA8964:, device type local memory mailbox, is online, record-oriented , >> >>     device, shareable, mailbox device. >> >> U >> >> Terminal VTA2621:, device type VT200 Series, is online, record-oriented device,  >> >>     carriage control.  >> >> L >> >> Again, as I wrote in my first response, the issue is file positioning.J >> >> If you are trying to pass an open file complete with record positionK >> >> to the subprocess, that's more difficult than simply passing a device K >> >> name.  And if you are trying to reposition the parent record position J >> >> to reflect records consumed by the child, that's more difficult thanP >> >> simply reading the next record from the mailbox/card reader/user keyboard. >> >> Q >> >> The important distinction is that record-oriented devices are non-seekable.  >> >8 >> > I goofed. I meant record-oriented files. Anyway,... >>J >> What would you mean by a non record-oriented file on VMS?  If you don'tE >> know then you have no business uttering the term "record-oriented" ' >> except in the context of a question.  > H > A file whose record format is undefined (see SET FILE/ATTR=RFM:UDF) orF > some really bizarre file from a non-VMS system are the only things I? > can think of, which is why I was confused in the first place.   7 A reasonable notion.  Just not the notion in play here.    > Hey, IH > didn't right the doc. But if you do SHOW DEVICE SYS$INPUT/FULL you areH > almost always going to get "record-oriented" in the output. (Actually, > I have yet to see otherwise.)    $ set verify $ @test  $ show device /full sys$Input   I Disk EISNER$DRA3:, device type 7 Member RAID 5, is online, mounted, file- O     oriented device, shareable, available to cluster, error logging is enabled.   O     Error count                    0    Operations completed           36027403 O     Owner process                 ""    Owner UIC                      [SYSTEM] O     Owner process ID        00000000    Dev Prot            S:RWPL,O:RWPL,G:R,W O     Reference count               39    Default buffer size                 512 O     Total blocks            17768448    Sectors per track                    64 O     Total cylinders            17352    Tracks per cylinder                  16   O     Volume label             "USER3"    Relative volume number                0 O     Cluster size                   3    Transaction count                    47 O     Free blocks             12015000    Maximum files allowed           2221056 O     Extend quantity                5    Mount count                           1 O     Mount status              System    Cache name      "_EISNER$DRA5:XQPCACHE" O     Extent cache size             64    Maximum blocks in extent cache  1201500 O     File ID cache size            64    Blocks currently in extent cache  89418 O     Quota cache size             162    Maximum buffers in FCP cache       1846 O     Volume owner UIC           [1,1]    Vol Prot    S:RWCD,O:RWCD,G:RWCD,W:RWCD   H   Volume Status:  ODS-2, subject to mount verification, do not unload onD       dismount, file high-water marking, write-back caching enabled.   $ exit  E When you use $ SHOW DEVICE SYS$INPUT /FULL from an interactive prompt A then the command will tell you that you are working from a record ! oriented device -- your terminal.   H When you use it from a command procedure you can get a different answer.  ) > I was trying to figure out what the DOC D > meant by non-record-oriented. Since the output of SHOW DEVICE/FULLH > SYS$INPUT has -- in my experience -- always indicated record-oriented,7 > I was confused as to what a NRO PPF w.r.t. SYS$INPUT.   C SYS$INPUT is ordinarily your command input.  If you are not running D under a CLI, it might be something else.  But if you're using $ SHOW7 DEVICE then it's clear that you're running under a CLI.   E >> > The manual says NON-record-oriented process-permanent FILES. The K >> > mailbox and terminal are record-oriented, so the table shouldn't apply L >> > to them. The disk is a device, not a file, so the table shouldn't apply >> > to it.  >>K >> Files on disk are not record-oriented in the relevant sense of the term.  >> >> I explained why.  >> >> Deal with it. > I > Fine. But this is a user manual and I wouldn't expect a typical user to G > have such sophistication as to correctly get the meaning of NRO PPF's H > in this case. I thought there had to be a simpler explanation. If I doG > SHOW DEVICE SYS$INPUT/FULL and it says that it's record-oriented, how @ > am I supposed to make the leap to the fact that it is actuallyH > non-record-oriented for this purpose? So I posted and you provided the > answer. Thank you.  G As I recall from when I first saw this documented in some release notes 0 a long time ago there were a number of examples.  C I'm not trying to defend the documentation.  I haven't read it in a 
 long time.   	John Briggs   ------------------------------  % Date: Wed, 14 Dec 2005 01:41:41 +0100 & From: Paul Sture <paul.sture@decus.ch>. Subject: Re: SHOW DEVICE: what should it show?, Message-ID: <409823F18q7ftU1@individual.net>   Bob Koehler wrote:` > In article <dnl1vg$k46$2@newslocal.mitre.org>, klewis@OMEGA.MITRE.ORG (Keith A. Lewis) writes: >  >>helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes in article <dnkrr8$gnj$1@online.de> dated Mon, 12 Dec 2005 21:59:04 +0000 (UTC): >>L >>>SHOW DEVICE shows a lot of devices, but not all.  For example, not NET*. I >>>Yes, with SHOW DEVICE NET you can see them, but not with no arguments.  >>>  >>>Bug or feature? >>* >>Feature.  A relatively new one, I think. >  > < >    Not new.  At least not new since VMS 2.5 on VAX 11/780. >   I Agreed. More than once I've tried to determine the name of a tape device     using:   $ SHOW DEVICE M   ) and been overwhelmed by MBA device names.   C Using a VT without a scroll buffer (pre VT520), CTRL-S if you were  6 lucky; bang it to a file or write a bit of DCL if not.   ------------------------------  % Date: Tue, 13 Dec 2005 16:35:01 -0800 # From: "Tom Linden" <tom@kednos.com> . Subject: Re: SHOW DEVICE: what should it show?( Message-ID: <ops1qzwnh3zgicya@hyrrokkin>  K On Wed, 14 Dec 2005 01:41:41 +0100, Paul Sture <paul.sture@decus.ch> wrote:    > Bob Koehler wrote:J >> In article <dnl1vg$k46$2@newslocal.mitre.org>, klewis@OMEGA.MITRE.ORG   >> (Keith A. Lewis) writes:  >>I >>> helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to   I >>> reply) writes in article <dnkrr8$gnj$1@online.de> dated Mon, 12 Dec    >>> 2005 21:59:04 +0000 (UTC): >>> I >>>> SHOW DEVICE shows a lot of devices, but not all.  For example, not   H >>>> NET*. Yes, with SHOW DEVICE NET you can see them, but not with no   >>>> arguments.  >>>> >>>> Bug or feature? >>> , >>> Feature.  A relatively new one, I think.? >>      Not new.  At least not new since VMS 2.5 on VAX 11/780.  >> > L > Agreed. More than once I've tried to determine the name of a tape device  
 >   using: >  > $ SHOW DEVICE M  > + > and been overwhelmed by MBA device names.  > F > Using a VT without a scroll buffer (pre VT520), CTRL-S if you were  8 > lucky; bang it to a file or write a bit of DCL if not.G What I usually do is to connect the console port to a W2K box and run   
 hyperterm.J Better access to SRM from VMS, I still think would be a nice thing, like   Tru64.   ------------------------------  + Date: Wed, 14 Dec 2005 00:50:28 +0000 (UTC) ? From: Graham Burley <burley.not-this@encompasserve-or-this.org>  Subject: Re: SRM memtest9 Message-ID: <439F6BEB.4B2013D0@encompasserve-or-this.org>    J.C. Roberts wrote:  >   H > If you background the memtest process and run show_status, it seems to > pass at least once?  >  > >>> memtest -rb -p 2 & > >>> show_status @ > ID        Program   Device   Pass   Hard/Soft   Written   ReadA > -------- --------- -------- ------ ----------- --------- ------ @ > 00000001      idle system       0      0    0         0      0@ > 0000004F   memtest memory       1      0    0         0      0  > When the test is actually doing something the Written and Read= counts should increase, when the test is complete the process > will disappear from show_status. If you Ctrl/T maybe you'd see, the memtest process in a "waiting on" state?  : > Using >>>kill_diags afterwards only locks up the system.  < You should be able to Ctrl/C it. My hunch is that when tests4 "stall" you need to INIT before trying further tests  . > When you guys use memtest, how do you do it?  @ I just had a play using your command (memtest -rb -p 2 &). I did@ a shutdown and memtest and the test "stalled", I did an INIT and the memtest ran ok.   
 >>>cat memory   B to see the "standard" memtest switches used when you run a >>>test   ------------------------------    Date: 13 Dec 2005 15:02:45 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) 9 Subject: RE: Timeout strategy: terminal vs Telnet drivers 3 Message-ID: <kgdrpuTWWQl2@eisner.encompasserve.org>   ` In article <JFEPKAPBPMDFDBOIANGDIEINHBAA.dallen@nist.gov>, "Dan Allen" <dallen@nist.gov> writes: >  >  >> -----Original Message----- G >> From: bill@triangle.cs.uofs.edu [mailto:bill@triangle.cs.uofs.edu]On  >> Behalf Of Bill Gunshannon+ >> Sent: Tuesday, December 13, 2005 9:07 AM  >> To: Info-VAX@Mvb.Saic.Com< >> Subject: Re: Timeout strategy: terminal vs Telnet drivers >> >>6 >> In article <AmxMupoFC8Ku@eisner.encompasserve.org>,3 >> 	Kilgallen@SpamCop.net (Larry Kilgallen) writes: H >> > In article <1134454677.069668.171280@o13g2000cwo.googlegroups.com>, >> dooleys@snowy.net.au writes:  >> >< >> >> a telnet packet is sent for every keystroke (and echo) >> >@ >> > What better argument could there be for DECnet's SET HOST ? >>K >> Why?  First, it is not necessarily an accurate description of how telnet 2 >> works and second, some applications require it. >> >> bill  > Q > Exactly - the Telnet application may WRITE individual characters but the TCP/IP P > layer may or may not send them as individual packets. TTBOMK applications haveK > no control over that phase of the network exchange - and for good reason.   E The context was "keystroke" and "echo", thus unrelated to application  output.   @ Applications have control over character input when they specifyB the termination characters they want honored.  VMS SET HOST passesC that information on to the system where the user is located and the B transmission is made only when the requirements of the applicationD have been satisfied.  "Tell me after every character" is a possible,? but rarely useful, possible set of requirements, even for TECO.    ------------------------------  + Date: Wed, 14 Dec 2005 03:12:56 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) 9 Subject: Re: Timeout strategy: terminal vs Telnet drivers ( Message-ID: <dno2jo$t7e$1@pcls4.std.com>  = koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:   c >In article <dnlg6n$5eo$1@pcls4.std.com>, moroney@world.std.spaamtrap.com (Michael Moroney) writes:  >>  K >> A VMS QIO READVBLK with IO$M_TIMED specified, and time (P3 parameter, I  I >> think) equal to 0 will give you whatever is in the typahead buffer (up I >> to the smaller of the buffer size/the typeahead size SYSGEN parameter) E >> and not wait at all for anything further.  This is how to get the  % >> nonblocking behavior through $QIO.   G >   I think somebody already stated that UCX's BGDRIVER doesn't support D >   IO$M_TIMED.  I assumed they knew what they were talking about as< >   I haven't used it the days when DEC still called it UCX.  F BGDRIVER should be irrelevant here.  The $QIO is processed by TTDRIVERF which likely handles the IO$M_TIMED modifier itself.  If not, the nextC layer is TNDRIVER, which is the "terminal" driver that talks to the E TCP socket.  BGDRIVER, the next layer should work in the usual way of  maintaining the TCP connection.    ------------------------------   End of INFO-VAX 2005.694 ************************