INFO-VAX Tue, 25 Sep 2007 Volume 2007 : Issue 523 Contents: Re: ba356 box Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Re: despair Re: Impact VMS version on CPU speed Re: Little Alpha I learned on is up... Re: Lock problem with SAMBA/NMBD Re: Lock problem with SAMBA/NMBD Re: MySQL failover on a OpenVMS Cluster Re: Re-routing MX bounce mails Re: Re-routing MX bounce mails Re: UK Based Alpha Station For Sale. Re: UK Based Alpha Station For Sale. Updated stunnel available for alpha and Itanium ---------------------------------------------------------------------- Date: Mon, 24 Sep 2007 18:56:38 -0500 From: David J Dachtera Subject: Re: ba356 box Message-ID: <46F84EB6.A3B1DBC7@spam.comcast.net> "rexdale1050@yahoo.ca" wrote: > > i am running open vms 7 .3 on alphaserver ds 10 using ds -ba356-rc > raid tower . can add an extra ds- ba356 to take over if original ds- > ba356 fails? > what hardware / software need > thank you for you help > neil Um, well, a BA356 is not a "RAID" tower, it's a SCSI disk "shelf", and provides device Id.'s 0 through 6. It may be possible to connect it to a RAID card in your computer's PCI backplane. If your RAID card supports dual SCSI busses it may be possible to configure it for redundancy using two disk shelves. Hope this helps... -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Mon, 24 Sep 2007 18:50:00 -0500 From: David J Dachtera Subject: Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Message-ID: <46F84D28.8043467D@spam.comcast.net> Gremlin wrote: > > Hi Richard > > Actually, over the last two years or so I have had many "robust" > discussions with them - but their final position is always a combination > of "HP isn't implementing the stack correctly" and "VMS - who cares?" > There answer to this issue is the former and they won't entertain any > further discussion about it.... Someone mentioned that the relevant RFC is dates sometime earlier this year. So, we MAY see the associated functionality in UCX sometime AFTER this time in 2009. ...but don't count on it. ...oh yeah: and if you really, REALLY, -REALLY- want it, make sure you can get The-VMS-Customer-Who-Shall-Not-Be-Named on your side. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Mon, 24 Sep 2007 18:53:29 -0500 From: David J Dachtera Subject: Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Message-ID: <46F84DF9.BF11ED54@spam.comcast.net> Gremlin wrote: > > Hi all > > Using OpenVMS 8.2 with HP TCP/IP v5.5 ECO 1. I am using the SmartFTP > client to get files to/from the box (FileZilla doesn't want to work) and > SmartFTP works OK except every file transferred incurs a lot ov > overhead throiugh directory re-listings. A short log file listing follows: > > Using SmartFTP v2.5.1006 to transfer html files to an OpenVMS server, I > get the following set of actions showing in the log: > [19:47:43] TYPE A > [19:47:43] 200 TYPE set to ASCII. > [19:47:43] PASV > [19:47:43] 227 Entering Passive Mode (x,y,z,200,238,55) > [19:47:43] Opening data connection to x.y.z.200 Port: 60983 > [19:47:43] LIST > [19:47:43] 150 Opening data connection for HT_ROOT:[000000]*.*;* > (x.y.z.55,2377) > [19:47:43] 20803 bytes transferred. (86.8 KB/s) (234 ms) > [19:47:43] 226 LIST Directory transfer complete. > [19:47:43] File "home.html" not found in directory listing. > [19:47:43] TYPE I > [19:47:43] 200 TYPE set to IMAGE. > [19:47:43] PASV > [19:47:43] 227 Entering Passive Mode (x,y,z,200,238,56) > [19:47:43] Opening data connection to x.y.z.200 Port: 60984 > [19:47:43] STOR home.html > [19:47:43] 150 Opening data connection for HT_ROOT:[000000]home.html; > (x.y.z.55,2378) > [19:47:44] 14448 bytes transferred. (129 KB/s) (109 ms) > [19:47:44] 226 Transfer complete. > [19:47:44] MDTM 20070923092635 home.html > [19:47:44] 502 MDTM is unimplemented. > [19:47:44] SIZE home.html > [19:47:44] 502 SIZE is unimplemented. > [19:47:44] SIZE not supported. > [19:47:44] Obtaining file information (size/date) from directory listing. > [19:47:44] Ignoring LIST options. > [19:47:44] TYPE A > [19:47:44] 200 TYPE set to ASCII. > [19:47:44] PASV > [19:47:44] 227 Entering Passive Mode (x,y,z,200,238,57) > [19:47:44] Opening data connection to x.y.z.200 Port: 60985 > [19:47:44] LIST > [19:47:44] 150 Opening data connection for HT_ROOT:[000000]*.*;* > (x.y.z.55,2379) > [19:47:44] 20905 bytes transferred. (86.8 KB/s) (235 ms) > [19:47:44] 226 LIST Directory transfer complete. > [19:47:44] File "home.html" not found in directory listing. > [19:47:44] Ignoring LIST options. > [19:47:44] PASV > [19:47:44] 227 Entering Passive Mode (x,y,z,200,238,59) > [19:47:44] Opening data connection to x.y.z.200 Port: 60987 > [19:47:44] LIST > [19:47:44] 150 Opening data connection for HT_ROOT:[000000]*.*;* > (x.y.z.55,2380) > [19:47:44] 20905 bytes transferred. (93.2 KB/s) (219 ms) > [19:47:44] 226 LIST Directory transfer complete. > [19:47:44] Transfer successful. > > SmartFTP *helpfully* say the problem is that OpenVMS doesn't support the > FTP SIZE or MDTM commands, so everythign has to be relisted after each > file. Agonising when transfering hundreds of them. > > Anyone have some suggestions about how I can implement SIZE and/or MDTM > in the FTP stack? The TCP/IP installation and management documentation > doesn't give me any hints. Is this a hobbyist machine? Could you try Multinet? Maybe inquire first over on vmsnet.networks.tcp-ip.multinet and see if they've implmented that RFC in a recent ECO. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Tue, 25 Sep 2007 10:36:23 +1000 From: Gremlin Subject: Re: Can't use SIZE and MDTM commands in OpenVMS FTP server Message-ID: <13fgm091cd57jcb@corp.supernews.com> Yes, it is a hobbyist machine and this is just my way of *not* running SAMBA on it - trying to reduce the security risks by not having a "live" link to the remainder of the internal network that can be exploited by MS type flaws.... David J Dachtera wrote: > Gremlin wrote: >> Hi all >> >> Using OpenVMS 8.2 with HP TCP/IP v5.5 ECO 1. I am using the SmartFTP >> client to get files to/from the box (FileZilla doesn't want to work) and >> SmartFTP works OK except every file transferred incurs a lot ov >> overhead throiugh directory re-listings. A short log file listing follows: >> >> Using SmartFTP v2.5.1006 to transfer html files to an OpenVMS server, I >> get the following set of actions showing in the log: >> [19:47:43] TYPE A >> [19:47:43] 200 TYPE set to ASCII. >> [19:47:43] PASV >> [19:47:43] 227 Entering Passive Mode (x,y,z,200,238,55) >> [19:47:43] Opening data connection to x.y.z.200 Port: 60983 >> [19:47:43] LIST >> [19:47:43] 150 Opening data connection for HT_ROOT:[000000]*.*;* >> (x.y.z.55,2377) >> [19:47:43] 20803 bytes transferred. (86.8 KB/s) (234 ms) >> [19:47:43] 226 LIST Directory transfer complete. >> [19:47:43] File "home.html" not found in directory listing. >> [19:47:43] TYPE I >> [19:47:43] 200 TYPE set to IMAGE. >> [19:47:43] PASV >> [19:47:43] 227 Entering Passive Mode (x,y,z,200,238,56) >> [19:47:43] Opening data connection to x.y.z.200 Port: 60984 >> [19:47:43] STOR home.html >> [19:47:43] 150 Opening data connection for HT_ROOT:[000000]home.html; >> (x.y.z.55,2378) >> [19:47:44] 14448 bytes transferred. (129 KB/s) (109 ms) >> [19:47:44] 226 Transfer complete. >> [19:47:44] MDTM 20070923092635 home.html >> [19:47:44] 502 MDTM is unimplemented. >> [19:47:44] SIZE home.html >> [19:47:44] 502 SIZE is unimplemented. >> [19:47:44] SIZE not supported. >> [19:47:44] Obtaining file information (size/date) from directory listing. >> [19:47:44] Ignoring LIST options. >> [19:47:44] TYPE A >> [19:47:44] 200 TYPE set to ASCII. >> [19:47:44] PASV >> [19:47:44] 227 Entering Passive Mode (x,y,z,200,238,57) >> [19:47:44] Opening data connection to x.y.z.200 Port: 60985 >> [19:47:44] LIST >> [19:47:44] 150 Opening data connection for HT_ROOT:[000000]*.*;* >> (x.y.z.55,2379) >> [19:47:44] 20905 bytes transferred. (86.8 KB/s) (235 ms) >> [19:47:44] 226 LIST Directory transfer complete. >> [19:47:44] File "home.html" not found in directory listing. >> [19:47:44] Ignoring LIST options. >> [19:47:44] PASV >> [19:47:44] 227 Entering Passive Mode (x,y,z,200,238,59) >> [19:47:44] Opening data connection to x.y.z.200 Port: 60987 >> [19:47:44] LIST >> [19:47:44] 150 Opening data connection for HT_ROOT:[000000]*.*;* >> (x.y.z.55,2380) >> [19:47:44] 20905 bytes transferred. (93.2 KB/s) (219 ms) >> [19:47:44] 226 LIST Directory transfer complete. >> [19:47:44] Transfer successful. >> >> SmartFTP *helpfully* say the problem is that OpenVMS doesn't support the >> FTP SIZE or MDTM commands, so everythign has to be relisted after each >> file. Agonising when transfering hundreds of them. >> >> Anyone have some suggestions about how I can implement SIZE and/or MDTM >> in the FTP stack? The TCP/IP installation and management documentation >> doesn't give me any hints. > > Is this a hobbyist machine? Could you try Multinet? > > Maybe inquire first over on vmsnet.networks.tcp-ip.multinet and see if they've > implmented that RFC in a recent ECO. > ------------------------------ Date: 24 Sep 2007 20:20:50 GMT From: Bruce Bowler Subject: Re: despair Message-ID: On Mon, 24 Sep 2007 12:32:07 -0400, Richard B. Gilbert wrote: > Bruce Bowler wrote: >> On Mon, 24 Sep 2007 08:14:22 -0400, Richard B. Gilbert wrote: >> >> >>>Fortran was the second high level language (COBOL was first) and dates >> >> >> Fortran's genesis was 1954-1957 (I've seen references to both) and >> COBOL's was 1959. >> >> Bruce > > See http://www.legacyj.com/cobol/cobol_history.html > > It looks as if 1959 was the date for CODASYL COBOL. Grace Hopper was > working on the problem as early as 1952. (A fast look at the problem > did not settle whether the language(s) she worked on was COBOL like.) And this... http://www.princeton.edu/~ferguson/adw/programming_languages.shtml Which puts Grace developing a compiler in '51 (A-0, not COBOL), FORTRAN in '57, LISP and ALGOL in '58, and COBOL (not mentioning CODASYL) in '59 or this... http://www.levenez.com/lang/history.html has FORTRAN starting '54 with B-0 (they don't mention A-0, but one would presume it came before B-0 :-) listed as a predecessor of COBOL in '57. Bruce ------------------------------ Date: Mon, 24 Sep 2007 20:10:03 -0700 From: Neil Rieck Subject: Re: Impact VMS version on CPU speed Message-ID: <1190689803.195093.64270@g4g2000hsf.googlegroups.com> On Sep 20, 3:38 am, wim.verst...@ns.nl wrote: > On 19 sep, 15:47, John Reagan wrote: > > > > > > > wim.verst...@ns.nl wrote: > > > The SYSTEM_CHECK parameter has on both machines the value 0. Both > > > machines are used for production purposes and there are no compilers > > > installed. So i used a very small Pascal program for comparison: > > > If you want a longer Pascal compilation to reduce the noise level, you > > can try: > > > $ pascal /terminal=stat /noobject /envir=tmp.pen /align=vax - > > sys$Library:starlet.pas > > > That should take several seconds of CPU time and even more seconds of > > elapsed time (lots of I/O to read the source and write the environment > > file). It took 4 or 5 CPU seconds on my XP1000 along with my pretty > > slow SCSI disks. On our GS1280 with nice 'n fast fibre disks, it flew. > > > $ pascal/noobj/envir=tmp.pen/terminal=stat/align=vax - > > sys$Library:starlet.pas > > CPU time: 0.62 seconds > > Elapsed time: 1.74 seconds > > Pagefaults: 1104 > > I/O Count: 4019 > > Peak memory: 0 kilobytes > > Source lines: 49154 > > 0 bytes per source line > > 4756838 lines per CPU minute > > 1694965 lines per elapsed minute > > > -- > > John Reagan > > OpenVMS Pascal/Macro-32/COBOL Project Leader > > Hewlett-Packard Company > > John, I think you misunderstood me. Both ES40's are production > machines without compilers installed. So i compiled the HBPRM source > on a development machine and copied the executable to both the ES40's. > I ran that image 99 times in a row and the results are as follows: > > On the 7.1-2 machine the total run took : > Charged CPU time: 0 01:10:44.95 > Elapsed time: 0 01:10:47.66 > > Lowest number of ticks: 4005 > Highest number of ticks: 5055 > Average after 99 runs : 4287 ticks per run. > > And on the 7.2-1 machine: > Charged CPU time: 0 01:09:10.45 > Elapsed time: 0 01:09:13.23 > > Lowest number of ticks: 3826 > Highest number of ticks: 5052 > Average after 99 runs : 4192 ticks per run. > > So, the 7.2-1 machine is a bit faster. My conclusion is that in our > version of 7.2-1 there is indeed a DCL bug. But one we can live with. > Thanks for all your contributions.- Hide quoted text - > > - Show quoted text - I don't have access to a Pascal complier so have nothing to add to this thread other than this: I have done ad-hoc bench-marking in BASIC and believe that every release of OpenVMS is faster than the previous version. But here are two really important points: 1) Any OpenVMS system with more that one CPU (SMP) should be running OpenVMS-7.3-2 at the very minimum. Prior to 7.3-2 many OpenVMS internal processes erroneously serialized on a single internal flag called IOLOCK8; not a problem with a single CPU but this limited SMP system scalability (adding more CPU's should have improved system performance but sometimes didn't). 2) The extended file cache (XFC) in OpenVMS-7.3-2 can combine with unused RAM to make your system turn from a roller skate into a roller coaster. My DS20E sped up so much that we convinced management to allow us to add an addtional 2 Gig of memory (to our original 1 Gig) and our users were certain that the system speed up by a factor of 4x. Neil Rieck Kitchener/Waterloo/Cambridge, Ontario, Canada. http://www3.sympatico.ca/n.rieck/ ------------------------------ Date: Mon, 24 Sep 2007 14:57:46 -0400 From: "Richard B. Gilbert" Subject: Re: Little Alpha I learned on is up... Message-ID: <46F808AA.7080602@comcast.net> FrankS wrote: > On Sep 24, 10:59 am, "Paul Raulerson" wrote: > >>You appear to have no idea of the business or technical or logistical >>considerations of moving software to a new platform. > > > Quite the contrary, but you wouldn't know anything about my > background. In fact, I've been the lead (well, sole, really) > developer for a few large and small projects that involved porting > applications from IBM and Unisys systems to OpenVMS. > > One thing I know for certain is that if you have software running on > OperatingSystem I and want to port it to OperatingSystem V then you > don't go out and buy a development system on which OperatingSystem V > was never supported. > How about if you already HAVE a system where O/S V "works" even if it's not "supported"? On a tight budget I'd at least try with the system I had! If it doesn't work out I can get new/used hardware in a couple of days. ------------------------------ Date: Mon, 24 Sep 2007 20:11:55 +0200 From: Albrecht Schlosser Subject: Re: Lock problem with SAMBA/NMBD Message-ID: <48bls4-rp1.ln1@news.hus-software.de> VAXman- @SendSpamHere.ORG wrote: > In article , Albrecht Schlosser writes: >> >> Martin Vorlaender wrote: >> >>> Albrecht Schlosser wrote: >>>> Volker Halle wrote: >>>>> You can use the File-ID field in the F11B$S >>>>> resource name to find out about the filename involved. >>>> I can't see anything obvious :-( Can you tell me how to >>>> find the file-ID ? >>> $ MCR DFU SEARCH /FID= >> Thanks for this, but this is not my problem. >> >> The question is, how to find the _file-ID_ from the lock >> info. > > ??? The file-ID is part of the file serialization lock name. > > Here's a little tutorial to help you find the file ID from the LOCK or > the RESOURCE block and how to trace it back to the file name itself. I > don't know if this is what you are asking or not; however, it doesn't > hurt to know how to derive this stuff regardless. [ long explanation removed ] That was _exactly_ what I was asking for :-) Thanks very much for your long explanation! I wrote a small procedure dump.com: $ x = %x'p1' $ dump/head/block=end=0/id=('x',0,0) dkb300: $ sea/stat/noout x.txt "''p1'7324 42313146" (Thanks to Jur van der Burg, too, for dump/id). This showed me the names and number of locks (last line) of the locks in question. And the winner is: [SAMBA.VAR]LOCKS.DIR with (currently) 23,773 of 23,780 locks. That's a big step forward finding the problem with NMBD :-) Regards Albrecht ------------------------------ Date: Mon, 24 Sep 2007 18:30:15 GMT From: VAXman- @SendSpamHere.ORG Subject: Re: Lock problem with SAMBA/NMBD Message-ID: In article <48bls4-rp1.ln1@news.hus-software.de>, Albrecht Schlosser writes: > > >VAXman- @SendSpamHere.ORG wrote: >> In article , Albrecht Schlosser writes: >>> >>> Martin Vorlaender wrote: >>> >>>> Albrecht Schlosser wrote: >>>>> Volker Halle wrote: >>>>>> You can use the File-ID field in the F11B$S >>>>>> resource name to find out about the filename involved. >>>>> I can't see anything obvious :-( Can you tell me how to >>>>> find the file-ID ? >>>> $ MCR DFU SEARCH /FID= >>> Thanks for this, but this is not my problem. >>> >>> The question is, how to find the _file-ID_ from the lock >>> info. >> >> ??? The file-ID is part of the file serialization lock name. >> >> Here's a little tutorial to help you find the file ID from the LOCK or >> the RESOURCE block and how to trace it back to the file name itself. I >> don't know if this is what you are asking or not; however, it doesn't >> hurt to know how to derive this stuff regardless. > >[ long explanation removed ] > >That was _exactly_ what I was asking for :-) Great. Finally some real VMS discussion here on c.o.v. >Thanks very much for your long explanation! > >I wrote a small procedure dump.com: > >$ x = %x'p1' >$ dump/head/block=end=0/id=('x',0,0) dkb300: >$ sea/stat/noout x.txt "''p1'7324 42313146" > >(Thanks to Jur van der Burg, too, for dump/id). I have to thank him too. I don't know how the /ID qualifier escaped under my radar all these years. I suppose I'll need to read the release notes a little more carefully. -- VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM "Well my son, life is like a beanstalk, isn't it?" http://tmesis.com/drat.html ------------------------------ Date: Mon, 24 Sep 2007 12:39:10 -0700 From: Keith Cayemberg Subject: Re: MySQL failover on a OpenVMS Cluster Message-ID: <1190662750.891410.177270@d55g2000hsg.googlegroups.com> On Sep 22, 1:55 am, "Jeffrey H. Coffield" wrote: > I now have to support MySQL on a OpenVMS cluster and need to have some > sort of way to switch it to another node in case the one it is running > on fails. It appears you can't have a copy on each node using the same > database files so I have thought of four possible ways. > > 1. Trigger some .COM file when a cluster transitions occurs that starts > MySQL on the node the cluster alias goes to. > > 2. Set up a replication slave on each other member but that means I > would have to set up some way to sync up the master when it came back > for writes to the database that occurred when the master was down. > > 3. Set up a MySQL cluster. I have only read about this and don't know > how stable this is. > > 4. Notify enough people that something has failed and that manual action > is needed. This is a new Itanium cluster so the most likely time the > master database would be down is when I am doing some sort of upgrade > that required more than just a quick reboot. > > As a related question, what is the best way to have a program launched > on a cluster transition. I currently have a program that needs a > constant tcpip connection so it gets an error when the cluster alias > fails over. > > Jeff Coffield Hi Jeff, please be careful not to over-simplify the steps taken to perform a fail-over. I assume the primary reason MySQL should not access the same DB files from two cluster members, is that MySQL does not use the OpenVMS DLM (Distributed Lock Manager). The DLM services used by cluster-aware databases or Record Management Services enables applications to independently access their data concurrently from two or more nodes of a cluster without causing data corruption. It is not clear to me that MySQL (Sybase also ignored the DLM at it's peril) is capable to otherwise recognize all circumstances where simultaneous access from multiple nodes would corrupt the database. This means you should provide your own method to insure a clean fail-over and single- node access locking that would otherwise be provided by a cluster- aware layer utilizing the DLM or some analog. The LCK package looks like a good tool to help achieve a clean fail- over, but there are a number of potential "gotchas" to be considered and handled with regard to performing a fail-over switch and holding it. 1) are you recognizing reliably all forms of a "permanently dead" vs. "non-responsive" vs. "temporarily non-responsive" vs. "dead now but soon recovered" node before deciding to switch nodes? 2) when deciding to perform a fail-over, how are you going to be 100% sure the other node is blocked from re-attaching to the the database files. Note, that this includes understanding all possible modes of connecting to the database from the old node, and understanding whether the blocking method is effective for all connection modes at all points of time during a reboot and loading of it's system software and layered products. Please also realise that your method of blocking access from the formerly killed node may for a colleague be indistinguishable from an incorrectly configured system. They may correct the condition, thus removing your block, and unintentionally corrupt your database. 3) before allowing DB access on the fail-over system it may be necessary to determine the state of the database. However, this is usually something any well engineered database needs to do anyway, since the situation is similar to that of simply pulling the plug on a stand-alone system, and restarting it. Normally, the database monitor must rollback unfinished transactions. However, to do this it needs access to some non-volatile journal of the transactions that were unfinished. It is important that the fail-over node will be able to locate and access the journal. It is unknown to me whether with MySQL you also need to recognize, wait for, or even initiate a successful rollback of the former node's transactions before or after connecting to the DB from the fail-over node. Here it may be interesting you to know that Oracle Rdb (I believe now also Oracle Classic with RAC) is fully cluster-aware also in this sense, the parallel running Monitor will recognize when another node has gone dead and will automatically perform the rollback of all it's transactions while continuing to provide local node access of the DB to all users. When the other node returns to the cluster, Rdb will then reliably allow the returned node to access the DB automatically. The points given above are only a simplified discourse of the possible situations that can arise in a two-node cluster, and which are also automatically handled by OpenVMS cluster-aware applications. Now consider how much more complicated the coordination of cases can become when handling larger multi-node clusters potentially having competing error conditions at different system software layers. Then it becomes clearer why when another product or company like IBM, Oracle, Tru/64, OCFS2 or the Redhat DLM are trying to attain some subset of VMS cluster functionality, they just copy the OpenVMS DLM design (in some cases even using some of the same VMS code). Correctly engineering a fully-shared cluster is one of the more difficult tasks ever to be attempted in informatics, and OpenVMS remains a unique masterpiece in this regard. The other clustering solutions are only simplified subsets that only offer their services to some subset of the system's processes, and then usually only for file access coordination. OpenVMS has clustering fully integrated in the OS, the cluster software layers are one of the first images loaded in a booting system (regardless of whether you have a cluster license), and OpenVMS provides cluster-wide access for nearly all OS objects, not just for some small subset of files on it's file system. Clustering capability is so integral to OpenVMS OS architecture that I can (and have) used a DCL script to watch processes starting on another booting cluster node before that node had even starting loading TCP/IP or DECnet. Cheers! Keith Cayemberg Consultant Wipro Technologies Sometimes when addressing "Big O" problems, I at least initially provide more questions than answers. ------------------------------ Date: Mon, 24 Sep 2007 19:54:05 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: Re-routing MX bounce mails Message-ID: > Anyone? I can't tell you how to fix it, but because spammers almost always forge real domain names in the From: address these days, this is something you really have to fix somehow. The formerly acceptable system of accepting all email and later bouncing undeliverables won't work any longer, and doing this frequently gets your site blacklisted as a spam source. Sorry that this probably won't help you at all. ------------------------------ Date: Tue, 25 Sep 2007 03:27:59 GMT From: Michael Austrin Subject: Re: Re-routing MX bounce mails Message-ID: <6e7ff418eed1ba6f8da67ae614d252ac$1@somehostoutintheEther.com> Michael Moroney wrote: >> Anyone? > I can't tell you how to fix it, but because spammers almost always forge > real domain names in the From: address these days, this is something you > really have to fix somehow. The formerly acceptable system of accepting > all email and later bouncing undeliverables won't work any longer, and > doing this frequently gets your site blacklisted as a spam source. > Sorry that this probably won't help you at all. I found a DCL spam killer - I modified it greatly... First the TCPIP ques are stopped. Another proc runs every 3 minutes that first uses a perl script to delete all non-user emails. I have a list of acceptable usernames and it goes through the xxxxxxxxxxx_TCPIP$SMTP-[pid].TCPIP_ file and determines the target user. If it is not in the list - it is deleted. Then the jobs are removed from the TCPIP ques and there is a manual process to delete any remaining unwanted mail after it is moved to a "hold" directory. Maybe one of these days I will get this upgraded to a later version of TCPIP that can do this itself... I also outright reject email coming from Russia or APAC. This was extremely helpful a few months ago when my 255Mhz Alpha 2100 processed more than 80K in one weekend. Michael Austin ------------------------------ Date: Mon, 24 Sep 2007 18:51:41 -0500 From: David J Dachtera Subject: Re: UK Based Alpha Station For Sale. Message-ID: <46F84D8D.812C3A7E@spam.comcast.net> Number8 wrote: > [snip] > Anyways, what's done is done, and if anyone is interested in it, the > url is > http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&rd=1&item=320162038276&ssPageName=STRK:MESE:IT&ih=011 > > As they say, thanks for looking. Well, I was smiling until I read the part about local pick-up only. My passport expired in 1994 and I'm not currently planning a trip to the UK anytime soon. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Mon, 24 Sep 2007 23:52:01 -0500 From: "Schroeder, AJ" Subject: Re: UK Based Alpha Station For Sale. Message-ID: <46f893bd$0$26349$4c368faf@roadrunner.com> David J Dachtera wrote: > Number8 wrote: >> [snip] >> Anyways, what's done is done, and if anyone is interested in it, the >> url is >> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&rd=1&item=320162038276&ssPageName=STRK:MESE:IT&ih=011 >> >> As they say, thanks for looking. > > Well, I was smiling until I read the part about local pick-up only. > > My passport expired in 1994 and I'm not currently planning a trip to > the UK anytime soon. If it's a 433 you want then I have a couple laying around. They are both 433au models. They both have a CD-ROM, a couple 9.1 Gb hard drives, I'll even throw in a DAT drive if you want. And I'm in the USA as well. ------------------------------ Date: Tue, 25 Sep 2007 08:06:05 +0800 From: "Richard Maher" Subject: Updated stunnel available for alpha and Itanium Message-ID: Hi, In case you haven't seen it, the following story at openvms.org discusses the VMS availability of the latest version of STUNNEL: - http://www.openvms.org/stories.php?story=07/09/21/0829898 Just for the record (and for those of you with the Tier3 Hobbyist kit) the Java Applet code provided is more than capable of talking to your Tier3 Application servers over the Web while being front-ended by STUNNEL. For those of you with dynamic client IP addresses or the IPsec deprived, this might be something you wish to pursue. If you are connecting to the Tier3 DEMO Queue Lookup example via STUNNEL then you simply need to edit t3$examples:cornucopiae.html and change the Applet Parameter SSL_REQD to the value "Y". The generic Java code (and thanks for your help again Arne) can be found at t3$examples:Tier3Socket.class. Cheers Richard Maher PS. If you wish to connect to a "self-certifying" server via this code then (thanks to one of the many outstanding contributions and help from Graham Burley) it was established that you'll need to register the server's certificate in the client's local Trust Store using the Java Keytool. For example: - keytool -importcert -alias manson -file d:\temp\manson.cer -keystore ..\lib\security\cacerts ------------------------------ End of INFO-VAX 2007.523 ************************