INFO-VAX Fri, 26 Jan 2007 Volume 2007 : Issue 52 Contents: Re: Delayed telnet connection How to catch ots$movc3 error in a pascal user process Re: How to get the codec from a file ? Re: Quorum disk removal Re: RMS Journaling Performance Re: SAS drops vms itanium port but not HP unix! ---------------------------------------------------------------------- Date: 26 Jan 2007 06:26:07 -0800 From: "RLFitch" Subject: Re: Delayed telnet connection Message-ID: <1169821567.276904.327350@h3g2000cwc.googlegroups.com> Changes made: --added entry to PC HOSTS file and --TCPIP SET SERVICE TELNET/LOG_OPTION=NOADDRESS Some improvement. Thanks to all! Ransom Fitch On Jan 24, 4:41 pm, "RLFitch" wrote: > When I attempt a telnet connection to my (local) server from a PC, I > can see the connection begin but then there's a 1-2 minute delay until > the login prompt is displayed. I'm not sure where to look for > troubleshooting, but I suspect the server DNS settings might be the > problem. I experience the same delay when I attempt a connection at > a remote site back to my server. This also happens while making FTP > connections. Using the same PC software I can make a telnet > connection to a remote server with no delay at all. Any suggestions? > > Info: > Alpha running OVMS 7.3-1 (Hobbyist) > PC running WinXP > > Thanks, > Ransom Fitchwww.rlfitch.com ------------------------------ Date: 26 Jan 2007 10:18:14 -0800 From: fkburrie@gmail.com Subject: How to catch ots$movc3 error in a pascal user process Message-ID: <1169835494.350458.35880@v33g2000cwv.googlegroups.com> Ls, One of my pascal programs crashed during ots$movc3 (Openvms 7.3.1) with pageread error (PAGRDERR) and io status 01F4 (=parity error?). Addressed memory is mapped to disk. According to HELP /message pagrderr an ana /disk /read_check has been performed without any significant result. The problem is that ots$movc3 does not return a result value that can be catched by the calling program. Someone has a clue of what could be the underlying problem and how I can catch the error? Thanks, Frank Burrie. ------------------------------ Date: Fri, 26 Jan 2007 12:26:07 -0500 From: Stephen Hoffman Subject: Re: How to get the codec from a file ? Message-ID: Klaus-D. Bohn wrote: > Sorry, but this is not enough for me. There's no reliable means to do what you want on OpenVMS nor on various other systems, and anything that you see is likely involving some guesswork. And anything you find will involve somebody porting over pieces -- there's nothing to do this in the base operating system. There's no storage of information containing what you want. You can port over a tool that examines the contents of the file and makes a guess as to the contents and the character set based on any sort of heuristic you might choose, but OpenVMS itself does not do that; there is no metadata on OpenVMS on the file data contents, save that created and managed by applications, and that necessary for RMS to access the data for the various file and record structures. I recall tools around that have guessed the language based on the tuples, and these could be brought to bear here. I recall tools that look at the file extension and make a guess. I recall tools that look at the file organization and make a guess. You can add transfer type "metadata" onto the anchor tags, too. But there is no means to do this that does not also involve some guessing, and there is no integrated means within OpenVMS to do what you want. (I've certainly used combinations of file extensions, RMS structure sniffing, and peeking at the characters in the data stream, as have others. This same basic area has nailed Microsoft Windows and Microsoft applications on at least several occasions, with corrupted or with maliciously-constructed files.) It might well be a nice enhancement to add this into OpenVMS, but any solution that involves doing this right -- application cooperation and the storage and availability of metadata -- is long past implementation within OpenVMS or similarly established operating systems. ("The curse of an installed base.") It would have to be retrofit, and applications that don't have or don't yet have the (hypothetical) metadata support would have to have their file(s) specifically managed somehow. If you do create or port over one or more heuristics, you can run it on a directory and tag the files with an ACE. Your tool to retrieve this can then look for the tag and skip the heuristics and related processing when the ACE is found. Applications using XML or such have a far better shot at this, as character encoding and language can be incorporated within the XML structures. xml:lang="en-US", and most any reasonable XML libraries can sort out the UTF encodings for you. OpenVMS doesn't support and doesn't know from this metadata (codecs, as you've referred to them), nor do OpenVMS applications. Hoff -- PS: I'd be quite willing to create and provide you with tools to sniff the RMS structures, and tools that maintain a metadata store on an ACL or elsewhere, if this is a production requirement. I've implemented this before. It can range from simple to fairly involved depending on the requirements, but the task itself is not overly difficult. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: Fri, 26 Jan 2007 12:36:36 -0500 From: Stephen Hoffman Subject: Re: Quorum disk removal Message-ID: John wrote: > What if you want to change your quorum disk from one device to another > device? Shut down the cluster (or the subset of nodes that are quorum watchers; the cluster member nodes with direct access to the quorum disk), reset the quorum disk system parameters (quorum_disk and maybe qdisk_votes), reboot with enough votes to reach quorum (without the quorum disk; either by booting enough nodes with enough votes, or by very carefully resetting the votes and expected_votes values in an environment that is not and cannot be partitioned), mount the (new) quorum disk, and off you go. If you have but one quorum watcher around or if the quorum disk is not otherwise connected to a multi-host bus, I'd suggest getting rid of the quorum disk and moving the votes into the host system. Ensure that votes and expected_votes are set per the FAQ. If you need to tweak the votes and expected_votes, make sure you set them back. Either by rebooting with new/old/correct values, or by clearing WRITESYSPARAMS back to zero if you're booting conversationally to temporarily change the quorum vote-related values. And since you're shuffling the quorum disk, quorum disks can be controller-shadowed so long as the controller will never show separate copies of the disk to two or more hosts. The quorum disk cannot be host-based volume shadowed. Having two different disks with the same disk name or the "same" disk contents is really bad in general. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: 26 Jan 2007 05:56:34 -0800 From: "Hein RMS van den Heuvel" Subject: Re: RMS Journaling Performance Message-ID: <1169819794.589110.305040@a75g2000cwd.googlegroups.com> A google for +rms +journaling +performance brings up a few useful pointers. (And +rms +journalling +performance, brings up more because I often misspell it that way ? :-) http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1063290 http://www.techiegroups.com/showthread.php?t=51998 http://h71000.www7.hp.com/wizard/wiz_2618.html Obviously Journalling will create overheard. I guess you are thinking AIJ, not RUJ or perhpas in combination with RUJ? RUJ is a must have for transactional integrity. Fairly cpu intense, lots of (little) IO to the same file over and over again. Like Solid State Disk. In use by major Stock exchanges. Little or no complaints. AIJ has low CPU overhead, but high IO volume. It's all about writing lots of data to large disks. You may adapt your tuning strategy some for journalling, picking your data bucket size as small as possible without risking that performance. Hope this helps some already, Hein van den Heuvel HvdH Performance Consulting. On Jan 26, 5:42 am, "Hal Kuff" wrote: > Hi, anyone have some comments about the performance impact of RMS Journaling > assu,ing no i/o throughput issues on the journaling disks? And anyone have > any issues with Journaling critical files in their experience? ------------------------------ Date: 26 Jan 2007 00:57:43 -0800 From: "Ian Miller" Subject: Re: SAS drops vms itanium port but not HP unix! Message-ID: <1169801863.550228.4750@k78g2000cwa.googlegroups.com> If you want to talk to SAS and/or HP about this then the appropriate people are listed on http://www.sas.com/partners/directory/hp/contacts.html ------------------------------ End of INFO-VAX 2007.052 ************************