INFO-VAX Sat, 31 May 2008 Volume 2008 : Issue 302 Contents: Re: Bletchley Park faces bleak future Re: CVS on VMS Re: CVS on VMS Re: CVS on VMS Re: Ever found a frayed DS10 internal SCSI cable? Re: Failsafe-IP Re: Failsafe-IP Re: Failsafe-IP Re: Failsafe-IP Re: Failsafe-IP Re: Failsafe-IP Re: Failsafe-IP Re: LBR function result codes still not available Re: LBR function result codes still not available Re: LBR function result codes still not available Noisy Fans DS10L Re: Noisy Fans DS10L Re: php with osu 3.10a and openvms 8.3 Re: Python on VMS status ? Re: Python on VMS status ? Re: Python on VMS status ? Re: Python on VMS status ? Re: TCP/IP ---------------------------------------------------------------------- Date: Fri, 30 May 2008 14:21:58 -0700 (PDT) From: IanMiller Subject: Re: Bletchley Park faces bleak future Message-ID: <1d5f9bde-0d51-45d7-b4a4-3c00a7bf8df2@a1g2000hsb.googlegroups.com> On 30 May, 17:26, "David Turner, Island Computers" wrote: > What would it take to keep it funded? > In PDS or $ > > -- > David B Turner > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > > Island Computers US Corp > PO Box 86 > Tybee GA 31328 > > Toll Free: 1-877 636 4332 x201, Mobile x251 > Email: dtur...@islandco.com > International & Local: (001)- 404-806-7749 > Fax: 912 786 8505 > Web:www.islandco.com > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D"IanMiller" <= g...@uk2.net> wrote in message > > news:5b357510-5781-44f7-9ce5-c60a6f78134e@k30g2000hse.googlegroups.com... > > > Off topic but hopefully of interest > > >http://www.bletchleypark.org.uk/news/docview.rhtm/516816 > > >http://www.bletchleypark.org.uk/ > > > Bletchley Park, which made huge contributions to early computing and > > shorterned WWII by years is in desperate need of money. > > Apparently the Dark Lord has turned down their request for funding > > which either demonstrates his cluelessness or he was not really > > asked. "The trust currently has to raise around =A31.5m a year just to maintain the site in its current condition." Mere pennies. Go buy something from their online shop. ------------------------------ Date: Fri, 30 May 2008 12:01:57 -0700 (PDT) From: marlow.andrew@googlemail.com Subject: Re: CVS on VMS Message-ID: <82a5b500-a7fb-4b95-87ab-7487bd95dfa2@d77g2000hsb.googlegroups.com> On 30 May, 17:24, "Richard Brodie" wrote: > wrote in message > > news:2f2d994a-c88c-40b0-a575-5a21861cd7a2@e39g2000hsf.googlegroups.com... > > > From googling I see that there have been attempts in the past to port > > CVS to VMS. AFAICS this development seems to have wound up in a > > backwater and I am not sure what the current state of play is. Can > > anyone advise please? > > CVS is a bit unfashionable these days. Yes, I know but where I am now they are way behind in the fashions.... > Folks are generally more > interested in ports of SVN or git. There are some reports of good > results with SVNkit. Where? Got any good URLs please? > The off the shelf alternative, is Perforce, I guess. Whilst I hear Perforce is very good there is no way they would exchange one proprietary solution for another. It would be very hard to argue the business benefit. > Otherwise, some assembly required. hmm. ------------------------------ Date: Fri, 30 May 2008 14:18:41 -0700 (PDT) From: IanMiller Subject: Re: CVS on VMS Message-ID: <8b0324e1-1cfa-49bf-ae5b-8145a6e80c40@m73g2000hsh.googlegroups.com> Searching everyones favourite vms news site I find the following about SVNkit http://www.openvms.org/stories.php?story=07/02/07/7115309 and also there is Mercurial source code control system http://www.openvms.org/stories.php?story=07/06/26/8837733 ------------------------------ Date: Fri, 30 May 2008 22:34:25 GMT From: =?ISO-8859-1?Q?Jan-Erik_S=F6derholm?= Subject: Re: CVS on VMS Message-ID: IanMiller wrote: > Searching everyones favourite vms news site I find the following about > SVNkit > http://www.openvms.org/stories.php?story=07/02/07/7115309 > > and also there is Mercurial source code control system > http://www.openvms.org/stories.php?story=07/06/26/8837733 Also note that Mercurial is preinstalled on the ready-built Python LD disk containers. See : http://vmspython.dyndns.org/ Jan-Erik. ------------------------------ Date: Fri, 30 May 2008 15:18:51 -0700 (PDT) From: dlzc Subject: Re: Ever found a frayed DS10 internal SCSI cable? Message-ID: <99608e24-5942-4b4f-94ca-84d9fd88f903@t12g2000prg.googlegroups.com> Dear tadamsmar: On May 30, 9:37=A0am, tadamsmar wrote: =2E.. > One theory is that it was ozone. =A0We have a one-pass > air handler and our ozone levels are relatively high > inside our building. =A0But I have not seen this with > other DS10s here. > > Or maybe some kind of vibration? Not likely ozone, unless the insulation was already damaged to unusability when the cable was fabricated. Heat or "solvents" wicking under the insulation would do this. SCSI drives that I have seen do get *very* hot, but not at the connector. Vibration would concentrate on the metal conductors as the stiffest part, and would not "overheat" the insulation. Likely vibration is what caused the functional failure of the drive... after the insulation failed and the conductors took all the vibration direct. I'd just check this cable again in a month and six months, to see if there is odd behavior there. If so, I would guess that it is the drive, and it may need a cooling fan directly on it. David A. Smith ------------------------------ Date: Fri, 30 May 2008 14:36:02 -0400 From: JF Mezei Subject: Re: Failsafe-IP Message-ID: <48404a26$0$7273$c3e8da3@news.astraweb.com> Gremlin wrote: > IP address. The desired result is that any request to the "cluster IP", > 192.168.0.10 would be served by any card in any host. Not possible. On ethernet, TCPIP uses ARP to find the ethernet address associated with a specific IP. When a device first communicates with say 10.0.0.17, it sends a broadcast "will 10.0.17 please stand up and identify itself". 10.0.0.17 then stands up, waves hands and sends a packet back to the requestor identifying itself. This gives the sender the ethernet address of 10.0.0.17 which it keeps in a cache. (TCPIP> SHOW ARP ) From then on, when the sender wants to communicate to 10.0.17 it finds its ethernet address locally in its own cache. After some inacivity timeout, the cache is invalidated and the next time the sender want to talk to 10.0.0.17, it needs to do the ARP thing again to re-establish the association between the IP and ethernet addresses. So an IP address can only be associated with one ethernet address at one time. The way the failover works is that when node B realises that node A has failed, node B will send what is called a gratuitous ARP response to everyone. everyone is then expected (but not required, not all devices support this) to update their ARP tables to have the new ethernet address (node B) associated with 10.0.0.17 Doing load balancing at the router level would also have interesting implications. Consider the case where node A sends a packet to the outside world, but when the outside world sends the ACK back, the router sends the ack to node B. This is one reason you had the failt tolerant vax, and nowadays that RTR software (is that atill alive, or was it declared mature ?) Put one very relaible machine up front which will distribute transactions to a series of nodes. This makes it possible to have a single external IP and do load balancing on many boxes. D ------------------------------ Date: Fri, 30 May 2008 14:25:40 -0700 (PDT) From: IanMiller Subject: Re: Failsafe-IP Message-ID: <02b9f6be-3476-401c-83ad-583bd865dde8@y38g2000hsy.googlegroups.com> On 30 May, 19:36, JF Mezei wrote: > Gremlin wrote: > > IP address. The desired result is that any request to the "cluster IP", > > 192.168.0.10 would be served by any card in any host. > > Not possible. > > On ethernet, TCPIP uses ARP to find the ethernet address associated with > a specific IP. > > When a device first communicates with say 10.0.0.17, it sends a > broadcast "will 10.0.17 please stand up and identify itself". 10.0.0.17 > then stands up, waves hands and sends a packet back to the requestor > identifying itself. This gives the sender the ethernet address of > 10.0.0.17 which it keeps in a cache. (TCPIP> SHOW ARP ) > > From then on, when the sender wants to communicate to 10.0.17 it finds > its ethernet address locally in its own cache. After some inacivity > timeout, the cache is invalidated and the next time the sender want to > talk to 10.0.0.17, it needs to do the ARP thing again to re-establish > the association between the IP and ethernet addresses. > > So an IP address can only be associated with one ethernet address at one > time. > > The way the failover works is that when node B realises that node A has > failed, node B will send what is called a gratuitous ARP response to > everyone. everyone is then expected (but not required, not all devices > support this) to update their ARP tables to have the new ethernet > address (node B) associated with 10.0.0.17 > > Doing load balancing at the router level would also have interesting > implications. Consider the case where node A sends a packet to the > outside world, but when the outside world sends the ACK back, the router > sends the ack to node B. > > This is one reason you had the failt tolerant vax, and nowadays that RTR > software (is that atill alive, or was it declared mature ?) > > Put one very relaible machine up front which will distribute > transactions to a series of nodes. This makes it possible to have a > single external IP and do load balancing on many boxes. > > D RTR is alive and well and living on Itanium systems. 5.1 came out last year. ------------------------------ Date: Sat, 31 May 2008 08:58:17 +1000 From: Gremlin Subject: Re: Failsafe-IP Message-ID: Thanks for all the suggestions. I understand the issues related to the NICs (thanks FrankS and Martin), but my need is slightly different. I am running a hobbyist cluster with three nodes. The router I use (can't afford a nice high-end load balancing or intelligent router) will only allow me to fordward port 80 to one internal IP address. Now, what I am hoping to do is provide some mechanism of having just one active cluster member provide a response to the port 80 request. I am not trying to do load balancing, I am just trying to service a port 80 request on any node that is active, without havining to update the router whenever a cluster member goes offline. I don't care if the ARP cache times out and take a while to find a new MAC to respond - it is only low-volume web serving and the latency is fine with me. Now, the excellent guide at http://h71000.www7.hp.com/openvms/journal/v2/articles/tcpip.html provided me with great additional information, including exactly what I wanted - the ClusterIP. However, trying to do a SHOW INTERFACE/CLUSTER faile on an OpenVMS 8.2 system with "parameter or qualifier CLUSTER no longer supported" So, I assume that means that cluster impersonation no longer exists? In the online Failsafe-IP documentation (not the one mentioned above), it isn't clear to me how to do *exactly* this - have a cluster impersonator IP. ------------------------------ Date: Fri, 30 May 2008 19:40:08 -0400 From: JF Mezei Subject: Re: Failsafe-IP Message-ID: <48409141$0$12331$c3e8da3@news.astraweb.com> Gremlin wrote: > Now, what I am hoping to do is provide some mechanism of having just one > active cluster member provide a response to the port 80 request. I am > not trying to do load balancing, I am just trying to service a port 80 > request on any node that is active, without havining to update the > router whenever a cluster member goes offline. In : SYS$STARTUP:TCPIP$SYSTARTUP.COM ! Standard parameters for any node $ SYSCONFIG = "$SYS$SYSTEM:TCPIP$SYSCONFIG.EXE" $ IFCONFIG = "$SYS$SYSTEM:TCPIP$IFCONFIG.EXE" $! $! create the cluster alias $ ifconfig we0 alias 10.0.0.50/16 ipmtu 1500 $ First node that executes this will get the lock to 10.0.0.50 and enable that interface. Subsequent nodes will configure it but put it on hold. When the "running" node fails, the next node that wants it will get it. This is the simple cluster failover functionality that existed before as a TCPIP> command. You can test failover with: ifconfig we0 -alias 10.0.0.50 This kills the alias completely and it then fails over to another node. You need to see if the other node's gratuitous ARP broadcast will be accepted by the router. If not, you will need to have a little script that runs every few minutes that connects to the router and clears its arp cache. ------------------------------ Date: Fri, 30 May 2008 18:40:30 -0500 From: David J Dachtera Subject: Re: Failsafe-IP Message-ID: <4840906E.77795EA@spam.comcast.net> Gremlin wrote: > > Hi All > > To begin with, thanks for the information given to previous questions - > no doubt that COV is a wealthy source of information and spirited debate! > > So, I have read the fine manuals for Failsafe-IP and, it appears to me, > that it is designed to allow me to have a VAX/Alpha/Itanic server that > contains two network cards to have the second card configured with an IP > address ready to take over if the first card fails. > > For example, if card A is 192.168.0.1 and card B is 192.168.0.2, then > card B can also be configured with a Failsafe-IP of 192.168.0.1 (and > card A could probably be confuigured with a Failsafe-IP of 192.168.0.2). > > From my reading, this appears to be the design extent of Failsafe-IP. > What I want to emulate is a cluster-wide single IP - almost as if each > card in each node in the cluster had a Failsafe-IP of say 192.168.0.10 > and that the Failsafe-IP could be active at the same time as the "real" > IP address. The desired result is that any request to the "cluster IP", > 192.168.0.10 would be served by any card in any host. > > Is this possible - my reading of the manual(s) suggests that the > Failsafe-IP is "dormant" until the "master" card fails. The point to remember is that UCX cannot detect a link failure above Layer 2. D.J.D. ------------------------------ Date: Sat, 31 May 2008 10:44:28 +1000 From: Gremlin Subject: Re: Failsafe-IP Message-ID: <8-KdnRKTEKjwAt3VnZ2dnUVZ_tHinZ2d@posted.internode> JF Mezei wrote: > Gremlin wrote: > >> Now, what I am hoping to do is provide some mechanism of having just one >> active cluster member provide a response to the port 80 request. I am >> not trying to do load balancing, I am just trying to service a port 80 >> request on any node that is active, without havining to update the >> router whenever a cluster member goes offline. > > > In : SYS$STARTUP:TCPIP$SYSTARTUP.COM > > ! Standard parameters for any node > $ SYSCONFIG = "$SYS$SYSTEM:TCPIP$SYSCONFIG.EXE" > $ IFCONFIG = "$SYS$SYSTEM:TCPIP$IFCONFIG.EXE" > $! > $! create the cluster alias > $ ifconfig we0 alias 10.0.0.50/16 ipmtu 1500 > $ > So, if I understand this correctly now..... NODE-A NODE-B NODE-C Home 10.0.0.1 10.0.0.2 10.0.0.3 Alias 10.0.0.100 10.0.0.100 10.0.0.100 I can point the router to direct all port 80 traffic to 10.0.0.100 and whichever node has the "lock" active will service the request? I do *not* want to point the router to a home address, just a "common" alias address. ------------------------------ Date: Fri, 30 May 2008 21:15:47 -0400 From: JF Mezei Subject: Re: Failsafe-IP Message-ID: <4840a7ac$0$12295$c3e8da3@news.astraweb.com> Gremlin wrote: > > So, if I understand this correctly now..... > > NODE-A NODE-B NODE-C > Home 10.0.0.1 10.0.0.2 10.0.0.3 > Alias 10.0.0.100 10.0.0.100 10.0.0.100 > > I can point the router to direct all port 80 traffic to 10.0.0.100 and > whichever node has the "lock" active will service the request? I do > *not* want to point the router to a home address, just a "common" alias > address. You understand correctly. You point your router to 10.0.0.100 and it will hit one of the 3 nodes above. When you do an ifconfig -a you get something like: WE0: flags=c43 Standby addresses: inet 10.0.0.50 netmask ffff0000 broadcast 10.0.255.255 (Active on VELO WE0) So it even tells you which node currently has the alias active. ------------------------------ Date: Fri, 30 May 2008 14:23:40 -0400 From: JF Mezei Subject: Re: LBR function result codes still not available Message-ID: <484047a3$0$29460$c3e8da3@news.astraweb.com> IanMiller wrote: > lbrdef contains structures and constants needed to call the LBR$ > routines. Not the condition values returned. I did it myself, and I am sure I would have gotten help from C.O.V. on how to get the ana/sys stuff. $ type lbrshr_codes.h /** LIBSHR_CODES.H ** extracted VAX-VMS 7.2 from ** ANA/SYS READ SYS$LIBRARY:LBRSHR.EXE , SHOW SYMBOL/ALL ** and then massages with TPU, keeping only LBR$_ codes **/ #define LBR$_LIBOPN 0x0026907A #define LBR$_LKPNOTDON 0x00269072 #define LBR$_NOFILNAM 0x00269082 #define LBR$_NOHISTORY 0x00268403 #define LBR$_NOHLPLIBS 0x00269102 #define LBR$_NOHLPTXT 0x0026908A #define LBR$_NOMTCHFOU 0x00268838 #define LBR$_NORMAL 0x00268001 #define LBR$_NOTHLPLIB 0x00269092 #define LBR$_NOUPDHIST 0x00268808 #define LBR$_NULIDX 0x00268810 #define LBR$_OLDLIBRARY 0x00268019 #define LBR$_OLDMISMCH 0x00268818 #define LBR$_RECLNG 0x0026909A #define LBR$_RECTRUNC 0x00268820 #define LBR$_REFCNTZERO 0x002690A2 #define LBR$_RFAPASTEOF 0x002690AA #define LBR$_STILLKEYS 0x00268828 #define LBR$_TOOMNYARG 0x002690F2 #define LBR$_TOOMNYLIB 0x002690B2 #define LBR$_TYPMISMCH 0x00268830 #define LBR$_UPDURTRAV 0x002690BA #define LBR$_USRINPERR 0x002690FA #define LBR$_WRITEERR 0x002690D2 #define LBR$_ALLWRNGBLK 0x00269002 #define LBR$_BADPARAM 0x002690C2 #define LBR$_DUPKEY 0x0026900A #define LBR$_EMPTYHIST 0x0026840B #define LBR$_ENDTOPIC 0x00268848 #define LBR$_ERRCLOSE 0x00268840 #define LBR$_HDRTRUNC 0x00268800 #define LBR$_ILLCREOPT 0x0026901A #define LBR$_ILLCTL 0x00269012 #define LBR$_ILLFMT 0x0026902A #define LBR$_ILLFUNC 0x00269032 #define LBR$_ILLIDXNUM 0x00269022 #define LBR$_ILLINROU 0x002690EA #define LBR$_ILLOP 0x0026903A #define LBR$_ILLOUTROU 0x002690DA #define LBR$_ILLOUTWID 0x002690E2 #define LBR$_ILLTYP 0x00269042 #define LBR$_INTRNLERR 0x002690CA #define LBR$_INVKEY 0x0026904A #define LBR$_INVNAM 0x00269052 #define LBR$_INVRFA 0x0026905A #define LBR$_KEYINDEX 0x00268009 #define LBR$_KEYINS 0x00268011 #define LBR$_KEYNOTFND 0x00269062 #define LBR$_LIBNOTOPN 0x0026906A $ ------------------------------ Date: Fri, 30 May 2008 12:05:59 -0700 From: "Tom Linden" Subject: Re: LBR function result codes still not available Message-ID: On Fri, 30 May 2008 11:23:40 -0700, JF Mezei wrote: > IanMiller wrote: >> lbrdef contains structures and constants needed to call the LBR$ >> routines. Not the condition values returned. > > I did it myself, and I am sure I would have gotten help from C.O.V. on > how to get the ana/sys stuff. > > $ type lbrshr_codes.h > /** LIBSHR_CODES.H > ** extracted VAX-VMS 7.2 from > ** ANA/SYS READ SYS$LIBRARY:LBRSHR.EXE , SHOW SYMBOL/ALL > ** and then massages with TPU, keeping only LBR$_ codes > **/ > > #define LBR$_LIBOPN 0x0026907A > #define LBR$_LKPNOTDON 0x00269072 > #define LBR$_NOFILNAM 0x00269082 > #define LBR$_NOHISTORY 0x00268403 > #define LBR$_NOHLPLIBS 0x00269102 > #define LBR$_NOHLPTXT 0x0026908A > #define LBR$_NOMTCHFOU 0x00268838 > #define LBR$_NORMAL 0x00268001 > #define LBR$_NOTHLPLIB 0x00269092 > #define LBR$_NOUPDHIST 0x00268808 > #define LBR$_NULIDX 0x00268810 > #define LBR$_OLDLIBRARY 0x00268019 > #define LBR$_OLDMISMCH 0x00268818 > #define LBR$_RECLNG 0x0026909A > #define LBR$_RECTRUNC 0x00268820 > #define LBR$_REFCNTZERO 0x002690A2 > #define LBR$_RFAPASTEOF 0x002690AA > #define LBR$_STILLKEYS 0x00268828 > #define LBR$_TOOMNYARG 0x002690F2 > #define LBR$_TOOMNYLIB 0x002690B2 > #define LBR$_TYPMISMCH 0x00268830 > #define LBR$_UPDURTRAV 0x002690BA > #define LBR$_USRINPERR 0x002690FA > #define LBR$_WRITEERR 0x002690D2 > #define LBR$_ALLWRNGBLK 0x00269002 > #define LBR$_BADPARAM 0x002690C2 > #define LBR$_DUPKEY 0x0026900A > #define LBR$_EMPTYHIST 0x0026840B > #define LBR$_ENDTOPIC 0x00268848 > #define LBR$_ERRCLOSE 0x00268840 > #define LBR$_HDRTRUNC 0x00268800 > #define LBR$_ILLCREOPT 0x0026901A > #define LBR$_ILLCTL 0x00269012 > #define LBR$_ILLFMT 0x0026902A > #define LBR$_ILLFUNC 0x00269032 > #define LBR$_ILLIDXNUM 0x00269022 > #define LBR$_ILLINROU 0x002690EA > #define LBR$_ILLOP 0x0026903A > #define LBR$_ILLOUTROU 0x002690DA > #define LBR$_ILLOUTWID 0x002690E2 > #define LBR$_ILLTYP 0x00269042 > #define LBR$_INTRNLERR 0x002690CA > #define LBR$_INVKEY 0x0026904A > #define LBR$_INVNAM 0x00269052 > #define LBR$_INVRFA 0x0026905A > #define LBR$_KEYINDEX 0x00268009 > #define LBR$_KEYINS 0x00268011 > #define LBR$_KEYNOTFND 0x00269062 > #define LBR$_LIBNOTOPN 0x0026906A > $ > > OK, JF your next assignment, if you want to make it really useful is to take the couple of minutes to put it into SDL :-) -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 30 May 2008 20:39:54 -0400 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: LBR function result codes still not available Message-ID: <48409e5a$0$90276$14726298@news.sunsite.dk> IanMiller wrote: > lbrdef contains structures and constants needed to call the LBR$ > routines. Not the condition values returned. Correct - it don't but it should ! Arne ------------------------------ Date: Fri, 30 May 2008 13:31:26 -0700 From: "Tom Linden" Subject: Noisy Fans DS10L Message-ID: Just noticed this. Anybody else experience this? It also is pretty hot to the touch. Also being klutzy accidentally bumped the power cord enough to bring it down, now all the shadow sets are merging, Grrr. -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 30 May 2008 16:50:12 -0400 From: JF Mezei Subject: Re: Noisy Fans DS10L Message-ID: <48406999$0$7255$c3e8da3@news.astraweb.com> Tom Linden wrote: > Just noticed this. Anybody else experience this? It also is pretty hot > to the touch. Also being klutzy accidentally bumped the power cord enough > to bring it down, now all the shadow sets are merging, Grrr. What is your DS10L running temperature ? $ write sys$output f$getsyi("TEMPERATURE_VECTOR") FFFFFFFFFFFFFFFFFFFFFFFFFFFFFF2C $ a = %x2C $ show symbol a A = 44 Hex = 0000002C Octal = 00000000054 My other DS10L, when turned on, runs at about 39-42. The one above has 2 disks so it runs a little hotter. (of course, in winter it is a couple degrees cooler). You might wish to open your DS10L and verify there are no wires near the fans. If a wire gets close enough, it might be being "rubbed" by fan blades (very bad thing), and this makes it noisy. You can lower the temperature of the unit by a couple of degrees by cutting out the grills on the rear fan outlet, and putting a better "fence" after the CPU to better redirect the hot hair towards the back and prevent hot air from recirculating to the front. (the little plastic tab provided to "curve" the air flow is not airtight, and there is also a gap that allows hot air to flow towards the area where the PCI card is located. These two improvements will lower the temp by 2°C Removing the grills for the front fans also helps increase airflow. ------------------------------ Date: Fri, 30 May 2008 20:38:41 -0400 From: =?ISO-8859-1?Q?Arne_Vajh=F8j?= Subject: Re: php with osu 3.10a and openvms 8.3 Message-ID: <48409e11$0$90276$14726298@news.sunsite.dk> Alan Winston - SSRL Central Computing wrote: > In article , Chuck Aaron writes: >> Will PHP run on OpenVMS 8.3 with the OSU 3.10a >> webserver or will PHP require CSWS? > > You can download PHP from the CSWS website, install and run PHP. > MOD_PHP won't work with OSU, but you can set up a presentation script for > .PHP pages that runs the PHP interpreter. But still PHP 4.3.10 that is very old and not supported any more by PHP development ? Arne ------------------------------ Date: Fri, 30 May 2008 12:08:49 -0700 (PDT) From: marlow.andrew@googlemail.com Subject: Re: Python on VMS status ? Message-ID: On 30 May, 18:33, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon Clubley) wrote: > What is the current status of Python on VMS ? > > Is the existing port considered to be robust and production ready ? I didn't know there was a python port? Got any good URLs? > > Are there any plans for HP to do a Perl/PHP style official port ? I don't know of any official perl but we have perl on our VMS system. Not sure how it got there.... > > Thanks, > > Simon. > > -- > Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP > Microsoft: Bringing you 1980's technology to a 21st century world ------------------------------ Date: Fri, 30 May 2008 12:19:52 -0700 From: "Tom Linden" Subject: Re: Python on VMS status ? Message-ID: On Fri, 30 May 2008 12:08:49 -0700, wrote: > I didn't know there was a python port? Got any good URLs? The Python RTE has received a significant update - WSGI. In a nut-shell, WSGI attempts to standardise the API for, and interactions with a Web-server, of a Web application written using Python - the Python Web Server Gateway Interface v1.0, as described in PEP 333 http://www.python.org/dev/peps/pep-0333/ The WSGI code has been developed under the sponsorship of SysGroup http://www.sysgroup.fr/ and generously been made available to the wider WASD community. The kit may be found on the WASD download page http://wasd.vsm.com.au/wasd/ and the new PyRTE source directory at http://wasd.vsm.com.au/ht_root/src/python/ Jean-François Piéronne will follow-up this email with news of Python on VMS, further Python/Web applications, and developments based on the WSGI model. Regards, Mark. -- PL/I for OpenVMS www.kednos.com ------------------------------ Date: Fri, 30 May 2008 20:33:14 GMT From: =?ISO-8859-1?Q?Jan-Erik_S=F6derholm?= Subject: Re: Python on VMS status ? Message-ID: marlow.andrew@googlemail.com wrote: > On 30 May, 18:33, clubley@remove_me.eisner.decus.org-Earth.UFP (Simon > Clubley) wrote: >> What is the current status of Python on VMS ? >> >> Is the existing port considered to be robust and production ready ? > > I didn't know there was a python port? Got any good URLs? http://vmspython.dyndns.org/ I uses the two LD-disks. Just to download and mount as LDA drives. They are pre-installed and rady to run. Works like a charm, inkl the Rdb interface and a lot of VMS specific system functions. Jan-Erik. ------------------------------ Date: Sat, 31 May 2008 08:59:39 +0930 From: Mark Daniel Subject: Re: Python on VMS status ? Message-ID: It's always flattering to find yourself quoted :-) but this one would be better in context. Jean-François Piéronne is the energetic 'team' that provided and maintains the port of Python for VMS (amongst other things). His Python-for-VMS site is http://vmspython.dyndns.org/ a Python based wiki made available via a WASD http://wasd.vsm.com.au/ enabled Web site running on VMS. WASD has an efficient, persistent environment for hosting Python web applications. What Tom has posted here is an announcement made on the info-WASD mailing list about recent enhancements to that environment. Not purely about VMS Python :-) Tom Linden wrote: > On Fri, 30 May 2008 12:08:49 -0700, wrote: > >> I didn't know there was a python port? Got any good URLs? > > > The Python RTE has received a significant update - WSGI. > > In a nut-shell, WSGI attempts to standardise the API for, and > interactions with a Web-server, of a Web application written using > Python - the Python Web Server Gateway Interface v1.0, as described in > PEP 333 > > http://www.python.org/dev/peps/pep-0333/ > > The WSGI code has been developed under the sponsorship of SysGroup > > http://www.sysgroup.fr/ > > and generously been made available to the wider WASD community. > > The kit may be found on the WASD download page > > http://wasd.vsm.com.au/wasd/ > > and the new PyRTE source directory at > > http://wasd.vsm.com.au/ht_root/src/python/ > > Jean-François Piéronne will follow-up this email with news of Python on > VMS, further Python/Web applications, and developments based on the > WSGI model. > > Regards, Mark. -- There may be fairies at the bottom of the garden. There is no evidence for it, but you can't prove that there aren't any, so shouldn't we be agnostic with respect to fairies? [Richard Dawkins; from The Nullifidian (Dec 94)] ------------------------------ Date: Fri, 30 May 2008 11:37:49 -0700 (PDT) From: Bob Gezelter Subject: Re: TCP/IP Message-ID: <34fbe7b1-863b-4e66-a04e-cd0f499d869b@m3g2000hsc.googlegroups.com> On May 30, 6:53 am, "Paul D Ainsworth" wrote: > From: "Paul D Ainsworth" > Subject: TCP/IP > Date: 30 May 2008 11:40 > > I know this is probably not the best NG for a post like this, but I need an > urgent answer and this group probably has the largest number of subscribers. > > I'm working on a project that sets up a blocking TCP/IP connection from an > openvms alpha ada executable to a python script running on a windows 2000 > advanced server. > > The connection sets up fine without any problems. However, the python script > should send back a status report in response to a status request from the > openvms executable. > > It sends back the status report but this causes a stack dump when received > by the alpha: - > > high performance arithmetic trap > floating invalid operation > arithmetic trap, floating overflow > arithmetic trap, floating underflow > > Does anyone have any ideas or suggestions towards a fix? Paul, Without the code, it is difficult to guess what is wrong. Having worked with the IO system for many years, and having done TCP connections on more than a few occasions, my first guess would be that an incorrect value was passed on the call to TCP (either the socket library or direct QIO). Such a bad value can also cause faults to occur in the user-mode libraries (e.g., sockets). Without the sources, it is difficult to make a substantive comment. If you wish to speak privately, please let me know (contact information, including email is on my www site). - Bob Gezelter, http://www.rlgsc.com ------------------------------ End of INFO-VAX 2008.302 ************************