INFO-VAX Thu, 07 Jun 2007 Volume 2007 : Issue 310 Contents: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... Re: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... Re: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... ALPHA_V732_MASTER_ECO_LIST.txt Re: ALPHA_V732_MASTER_ECO_LIST.txt Comparative patch installation (was Re: ALPHA_V732_MASTER_ECO_LIST.txt) Re: Disabling scp and sftp on ssh-server? Re: FYI - Limit to the ACP Caches Re: FYI - Limit to the ACP Caches Global warming NOT causing more hurricanes UK scientists say Re: Global warming NOT causing more hurricanes UK scientists say Re: Global warming NOT causing more hurricanes UK scientists say Re: Global warming NOT causing more hurricanes UK scientists say Re: Global warming NOT causing more hurricanes UK scientists say HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: HOWTO: Ignoring wildly OT threads Re: Impact of an abnormally high RECNXINTERVAL Re: Impact of an abnormally high RECNXINTERVAL Re: Impact of an abnormally high RECNXINTERVAL Re: Not on latest Roadmap: OpenVMS VAX Version 8.x "under investigation" Re: Not on latest Roadmap: OpenVMS VAX Version 8.x "under investigation" Re: porting ignorance (was Re: Story Time) Re: porting ignorance (was Re: Story Time) Re: porting ignorance (was Re: Story Time) Re: porting ignorance (was Re: Story Time) Re: porting ignorance (was Re: Story Time) Re: porting ignorance (was Re: Story Time) Re: SECURITY.AUDIT$JOURNAL and ERRLOG.SYS Re: Story Time Re: Story Time Re: Story Time Re: Story Time Re: Thin wire Coax Ethernet on Alpha using RBLs interactively Re: using RBLs interactively Re: VMS Update Monday June 4th ---------------------------------------------------------------------- Date: Thu, 07 Jun 2007 07:42:14 -0700 From: Galen Subject: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... Message-ID: <1181227334.569969.214560@n4g2000hsb.googlegroups.com> We're running TCP/IP V5.3 (no TCP/IP ecos) on VMS V7.3-1. The system is not using BIND but is using TCPIP$HOSTS.DAT instead. On only one of several systems running this configuration TCPIP SHOW ROUTES gives: %TCPIP-E-ROUTEERROR, error accessing routes database (TCPIP$ROUTE) yet TCPIP SHOW ROUTES/PERMANENT works just fine. We tried creating a new TCPIP$ROUTE.DAT (TCPIP CREATE ROUTE) and re- entering our routes. We had to use SET ROUTE/PERMANENT or we got the same error as above. The original problem also persisted. Restarting TCP/IP did not help, nor did rebooting the system. Searching with google has turned up very little, none of which looked much like our case. (We aren't using BIND or DHCP which were involved in most of what I found.) My next step will be to remove TCP/IP and its database files, then completely reinstall it. Any other suggestions to try first? Thanks, Galen ------------------------------ Date: Thu, 07 Jun 2007 11:23:47 -0400 From: JF Mezei Subject: Re: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... Message-ID: Galen wrote: > On only one of several systems running this configuration TCPIP SHOW > ROUTES gives: > %TCPIP-E-ROUTEERROR, error accessing routes database (TCPIP$ROUTE) I recall having this once a long time ago. Do you have any definitions in the hosts database ? (does TCPIP> SET HOST pastry.chocolate.com/address=1.2.3.4 work ? If you do: SHOW HOST pastry.chocolate.com does it work ? of you do SHOW HOST does it show the defined hosts and then complain about the missing route file ? Have you also tried TCPIP> CREATE NETWORK ? You should have the following files : $ dir sys$system:tcpip*.dat* Directory SYS$COMMON:[SYSEXE] TCPIP$CONFIGURATION.DAT;1 TCPIP$HOST.DAT;1 TCPIP$NETWORK.DAT;1 TCPIP$PROXY.DAT;1 TCPIP$ROUTE.DAT;1 TCPIP$SERVICE.DAT;1 ------------------------------ Date: Thu, 07 Jun 2007 10:19:03 -0700 From: Galen Subject: Re: %TCPIP-E-ROUTEERROR from TCPIP SHOW ROUTES... Message-ID: <1181236743.446405.222600@p47g2000hsd.googlegroups.com> > I recall having this once a long time ago. > Have you also tried TCPIP> CREATE NETWORK ? > No effect (already had an empty TCPIP$NETWORK.DAT anyway) > You should have the following files : > > $ dir sys$system:tcpip*.dat* > > Directory SYS$COMMON:[SYSEXE] > > TCPIP$CONFIGURATION.DAT;1 TCPIP$HOST.DAT;1 > TCPIP$NETWORK.DAT;1 > TCPIP$PROXY.DAT;1 TCPIP$ROUTE.DAT;1 TCPIP$SERVICE.DAT;1 Got 'em all. > Do you have any definitions > in the hosts database ? > (does TCPIP> SET HOST pastry.chocolate.com/address=1.2.3.4 work ? > Yes... > If you do: > > SHOW HOST pastry.chocolate.com does it work ? > of you do SHOW HOST does it show the defined hosts and then complain > about the missing route file ? > $ TCPIP SHOW ROUTE %TCPIP-E-ROUTEERROR, error accessing routes database (TCPIP$ROUTE) $ TCPIP TCPIP>SHOW HOST ! ! List of hosts ! TCPIP>SHOW ROUTE ! ! List of routes, WITHOUT any error message! ! JF, I think you're onto the problem here. What's going on, if you know? ------------------------------ Date: Thu, 7 Jun 2007 08:44:34 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: ALPHA_V732_MASTER_ECO_LIST.txt Message-ID: Soon, VMS732_UPDATE-V1200 will appear. I recently installed UPDATE-V1100 and a few patches which came out after it. In ALPHA_V732_MASTER_ECO_LIST.txt one can read "At the current time, no further UPDATE kits are scheduled for OpenVMS V7.3-2". Since UPDATE-V1100 was released on 7-MAR-2007, it seems that the plans changed rather quickly. Anyway, the main question is: what is in this UPDATE patch which I don't already have. ALPHA_V732_MASTER_ECO_LIST.txt has two lists: KITS INCLUDED IN UPDATE KIT (with a note if there is a newer version of the patch than the one in the kit) and ECO KITS NOT INCLUDED IN UPDATE KIT The latter contains mostly the newer versions of the ECOs which are marked as superseded in the former. It would be nice to have a third list, namely ECO KITS NOT INCLUDED IN THE PREVIOUS UPDATE KIT since that is probably the thing most people are interested in. Basically, I have to go through the list in the section CONSOLIDATION OF PREVIOUSLY RELEASED V7.3-2 ECO KITS and check which of the ECOs are installed. I can't use PRODUCT because some of them will have been installed via an UPDATE patch, so I have to check the release notes. Actually, it would be enough if in the lists of ECOs, release notes and images there was a note indicating if the corresponding patch wasn't in the previous update kit. Can the patch maintainer at HP implement this relatively easy change? Does anyone have any code to automate this (i.e. from the contents of the release notes in SYS$HELP and the contents of ALPHA_V732_MASTER_ECO_LIST.txt determine which patches I don't yet have? ------------------------------ Date: Thu, 7 Jun 2007 09:18:58 -0400 From: norm.raphael@metso.com Subject: Re: ALPHA_V732_MASTER_ECO_LIST.txt Message-ID: Actually, this becomes moot, because after the UPDATE patch is release, it starts to become a prerequisite for new patches. Thus, even if one has installed all ECO's in the latest UPDATE ECO, one must perforce install the UPDATE ECO in order to have later-released ECO's install successfully. I once asked if the UPDATE ECO could check for the presence of installed ECO contained within and just either install the ones not yet present or (since if some we not there it may well be because the are unwanted/unneeded--non-category-one ECO's somehow get into these UPDATE ECO's), provide a switch to just mark the database so later ECO's don't fail; IOW trust the system manager--what a concept. The comment was that this was too hard to do. helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) wrote on 06/07/2007 04:44:34 AM: > Soon, VMS732_UPDATE-V1200 will appear. I recently installed > UPDATE-V1100 and a few patches which came out after it. In > ALPHA_V732_MASTER_ECO_LIST.txt one can read "At the current time, no > further UPDATE kits are scheduled for OpenVMS V7.3-2". Since > UPDATE-V1100 was released on 7-MAR-2007, it seems that the plans changed > rather quickly. > > Anyway, the main question is: what is in this UPDATE patch which I don't > already have. ALPHA_V732_MASTER_ECO_LIST.txt has two lists: > > KITS INCLUDED IN UPDATE KIT (with a note if there is a newer version > of the patch than the one in the kit) > > and > > ECO KITS NOT INCLUDED IN UPDATE KIT > > The latter contains mostly the newer versions of the ECOs which are > marked as superseded in the former. > > It would be nice to have a third list, namely > > ECO KITS NOT INCLUDED IN THE PREVIOUS UPDATE KIT > > since that is probably the thing most people are interested in. > > Basically, I have to go through the list in the section > > CONSOLIDATION OF PREVIOUSLY RELEASED V7.3-2 ECO KITS > > and check which of the ECOs are installed. I can't use PRODUCT because > some of them will have been installed via an UPDATE patch, so I have to > check the release notes. > > Actually, it would be enough if in the lists of ECOs, release notes and > images there was a note indicating if the corresponding patch wasn't in > the previous update kit. > > Can the patch maintainer at HP implement this relatively easy change? > > Does anyone have any code to automate this (i.e. from the contents of > the release notes in SYS$HELP and the contents of > ALPHA_V732_MASTER_ECO_LIST.txt determine which patches I don't yet have? > ------------------------------ Date: Thu, 07 Jun 2007 08:35:59 -0500 From: Ron Johnson Subject: Comparative patch installation (was Re: ALPHA_V732_MASTER_ECO_LIST.txt) Message-ID: <3RT9i.23183$gM1.5893@newsfe21.lga> You know... installing "patches" in Linux is a *lot* simpler. On 06/07/07 08:18, norm.raphael@metso.com wrote: > Actually, this becomes moot, because after the UPDATE patch is release, it > starts to become a prerequisite for new patches. > Thus, even if one has installed all ECO's in the latest UPDATE ECO, one > must perforce install the UPDATE ECO in order to > have later-released ECO's install successfully. > > I once asked if the UPDATE ECO could check for the presence of installed > ECO contained within and just either install the ones > not yet present or (since if some we not there it may well be because the > are unwanted/unneeded--non-category-one ECO's > somehow get into these UPDATE ECO's), provide a switch to just mark the > database so later ECO's don't fail; IOW trust the > system manager--what a concept. The comment was that this was too hard to > do. > > > helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) > wrote on 06/07/2007 04:44:34 AM: > >> Soon, VMS732_UPDATE-V1200 will appear. I recently installed >> UPDATE-V1100 and a few patches which came out after it. In >> ALPHA_V732_MASTER_ECO_LIST.txt one can read "At the current time, no >> further UPDATE kits are scheduled for OpenVMS V7.3-2". Since >> UPDATE-V1100 was released on 7-MAR-2007, it seems that the plans changed >> rather quickly. >> >> Anyway, the main question is: what is in this UPDATE patch which I don't >> already have. ALPHA_V732_MASTER_ECO_LIST.txt has two lists: >> >> KITS INCLUDED IN UPDATE KIT (with a note if there is a newer version >> of the patch than the one in the kit) >> >> and >> >> ECO KITS NOT INCLUDED IN UPDATE KIT >> >> The latter contains mostly the newer versions of the ECOs which are >> marked as superseded in the former. >> >> It would be nice to have a third list, namely >> >> ECO KITS NOT INCLUDED IN THE PREVIOUS UPDATE KIT >> >> since that is probably the thing most people are interested in. >> >> Basically, I have to go through the list in the section >> >> CONSOLIDATION OF PREVIOUSLY RELEASED V7.3-2 ECO KITS >> >> and check which of the ECOs are installed. I can't use PRODUCT because >> some of them will have been installed via an UPDATE patch, so I have to >> check the release notes. >> >> Actually, it would be enough if in the lists of ECOs, release notes and >> images there was a note indicating if the corresponding patch wasn't in >> the previous update kit. >> >> Can the patch maintainer at HP implement this relatively easy change? >> >> Does anyone have any code to automate this (i.e. from the contents of >> the release notes in SYS$HELP and the contents of >> ALPHA_V732_MASTER_ECO_LIST.txt determine which patches I don't yet have? >> > -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 7 Jun 2007 12:10:54 +0000 (UTC) From: Pekka Lylykorpi Subject: Re: Disabling scp and sftp on ssh-server? Message-ID: Richard Whalen wrote: : "Pekka Lylykorpi" wrote in message : news:f1v6am$4ja$1@news.cc.tut.fi... :>I have OpenVMS knowledge next to zero but I try to explain our :> problem with ssh2 on OpenVMS. :> :> We have an application on started directly (in LOGIN.COM) :> when users log in and when users quit that application they :> will be automatically logged out. They don't see any files :> on file system level (ftp is not granted). When they use :> telnet this works fine. :> :> But when ssh (Ssh Communications Tectia client) is used :> instead of telnet people can also use it's integrated file :> transfer and see and access files on file system level. Is :> there any mean to limit that access (to disable both sftp :> and scp)? : Comment out the line that begins : subsystem-sftp : in sys$sysdevice:[tcpip$ssh.ssh2]SSHD2_CONFIG. Thanks, this helped! Pekka ------------------------------ Date: Thu, 07 Jun 2007 15:45:33 -0000 From: Hein RMS van den Heuvel Subject: Re: FYI - Limit to the ACP Caches Message-ID: <1181231133.448517.259790@m36g2000hse.googlegroups.com> On Jun 6, 10:26 pm, daryljo...@att.net wrote: > Hello Everyone: > > This is FYI. > > Thanks for your recent call about the inability of Sysgen to detect > and report erroneous values for ACP_HRDCACHE in OpenVMS Alpha V8.3 > > The problem is actually worse than you pointed out. The internal > parameter is stored in a 16-bit word. Interesting / Timely topic. I've just been working on / involved with / a system where we were wondering how far to jack up HDR_CACHE. In that particular case it would be mostly a workaround untill we get fragmenation under control. How is this for bad fragmentation... a file with 9500 HEADERS, not just extents, headers full of extents. Alongside several active files with 100+ extention headers and a critical indexed file (100+ GB) with 1500 headers. The last one is mutli-key and the activity is likely to hit a large range of the file VBNs. Result: hdr cache attemopt rate at 8000/sec, fcp read rate approachin 2000/sec, high Kernel time, High Mpsync time. We bumped the cache to 30,000 (form 5k?) but with no immediate positive result. An outage was declared and the two worst offenders moved around some. Cheers, Hein. ------------------------------ Date: Thu, 07 Jun 2007 12:35:09 -0400 From: "Richard B. Gilbert" Subject: Re: FYI - Limit to the ACP Caches Message-ID: <466833BD.6000700@comcast.net> Hein RMS van den Heuvel wrote: > On Jun 6, 10:26 pm, daryljo...@att.net wrote: > >>Hello Everyone: >> >>This is FYI. >> >>Thanks for your recent call about the inability of Sysgen to detect >>and report erroneous values for ACP_HRDCACHE in OpenVMS Alpha V8.3 >> >>The problem is actually worse than you pointed out. The internal >>parameter is stored in a 16-bit word. > > > Interesting / Timely topic. > > I've just been working on / involved with / a system where we were > wondering how far to jack up HDR_CACHE. In that particular case it > would be mostly a workaround untill we get fragmenation under control. > > How is this for bad fragmentation... a file with 9500 HEADERS, not > just extents, headers full of extents. Alongside several active files > with 100+ extention headers and a critical indexed file (100+ GB) with > 1500 headers. The last one is mutli-key and the activity is likely to > hit a large range of the file VBNs. > > Result: hdr cache attemopt rate at 8000/sec, fcp read rate approachin > 2000/sec, high Kernel time, High Mpsync time. > We bumped the cache to 30,000 (form 5k?) but with no immediate > positive result. > An outage was declared and the two worst offenders moved around some. > > Cheers, > Hein. > > Seems to me that there must be an underlying problem or two! Application design? Allocating files two blocks at a time? Time for an image backup, disk-to-disk, and swap the disks? ------------------------------ Date: Thu, 07 Jun 2007 06:01:07 -0700 From: ultradwc@gmail.com Subject: Global warming NOT causing more hurricanes UK scientists say Message-ID: <1181221267.476748.239100@q66g2000hsg.googlegroups.com> why will not Andrew listen to his own English scientists? :) http://www.timesonline.co.uk/tol/news/world/europe/article1896266.ece ------------------------------ Date: Thu, 07 Jun 2007 07:31:30 -0700 From: Andrew Subject: Re: Global warming NOT causing more hurricanes UK scientists say Message-ID: <1181226690.295872.97130@w5g2000hsg.googlegroups.com> On 7 Jun, 14:01, ultra...@gmail.com wrote: > why will not Andrew listen to his own English scientists? :) > > http://www.timesonline.co.uk/tol/news/world/europe/article1896266.ece Quoting for the article "The findings are at variance with the conclusions in February of the Intergovernmental Panel on Climate Change, the United Nations organisation addressing global warming. The UN panel stopped short of blaming increased frequency of hurricanes on man-made temperature rises, but said it was "more likely than not" that greenhouse gas emissions had contributed to the greater intensity of cyclonic storms. The Swedish-led research team suggested that hurricane levels were normal, though they accepted "a future possibility" of higher sea temperatures contributing to more intense hurricanes.The researchers were unable to identify any direct link between increased hurricanes and rising sea level temperatures, beyond the requirement for a minimum temperature of 27C (81F) to be reached before a hurricane developed. " Yet again your lack of knowledge about basic geography lets you down, Sweden is not part of the UK nor is Puerto Rico or the US though you were once a colony. And the Swedish-led report differs only from the IPCC report in that the IPCC report said that it was more likely than not that greenhouse gas emissions had contributed to the greater intensity of cyclonic storms while the Swedish report accepted that there was a future possibility that cyclonic storms could be of greater intensity due to GW. Nor does the Swedish report question the reality of GW itself. All in all not your best attempt. D- Regards Andrew ------------------------------ Date: 7 Jun 2007 10:28:40 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: Global warming NOT causing more hurricanes UK scientists say Message-ID: In article <1181221267.476748.239100@q66g2000hsg.googlegroups.com>, ultradwc@gmail.com writes: > > why will not Andrew listen to his own English scientists? :) Why will you not stick to on-topic posts ? ------------------------------ Date: Thu, 07 Jun 2007 10:06:04 -0700 From: ultradwc@gmail.com Subject: Re: Global warming NOT causing more hurricanes UK scientists say Message-ID: <1181235964.394926.208870@h2g2000hsg.googlegroups.com> On Jun 7, 11:28 am, Kilgal...@SpamCop.net (Larry Kilgallen) wrote: > In article <1181221267.476748.239...@q66g2000hsg.googlegroups.com>, ultra...@gmail.com writes: > > > why will not Andrew listen to his own English scientists? :) > > Why will you not stick to on-topic posts ? I just want to help Andrew not look like a fool to his own countrymen ... :) ------------------------------ Date: 7 Jun 2007 17:08:30 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Global warming NOT causing more hurricanes UK scientists say Message-ID: <5cqsceF30troeU1@mid.individual.net> In article <1181235964.394926.208870@h2g2000hsg.googlegroups.com>, ultradwc@gmail.com writes: > On Jun 7, 11:28 am, Kilgal...@SpamCop.net (Larry Kilgallen) wrote: >> In article <1181221267.476748.239...@q66g2000hsg.googlegroups.com>, ultra...@gmail.com writes: >> >> > why will not Andrew listen to his own English scientists? :) >> >> Why will you not stick to on-topic posts ? > > I just want to help Andrew not look like a fool to his > own countrymen ... :) It would seem prudent for you to pay more attention to how you look to yours! bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Thu, 07 Jun 2007 08:34:02 -0500 From: Ron Johnson Subject: HOWTO: Ignoring wildly OT threads Message-ID: Any dedicated newsreader worth it's salt has an "Ignore thread" feature. Thunderbird 2.0 finally also has an "Ignore thread" in it's news function: Message->"Ignore Thread". The hotkey is "K". -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 07 Jun 2007 09:43:29 -0400 From: Paul Anderson Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: In article , Ron Johnson wrote: > Any dedicated newsreader worth it's salt has an "Ignore thread" > feature. I heard that doing that contributes to global warming. Paul -- Paul Anderson OpenVMS Engineering Hewlett-Packard Company ------------------------------ Date: 7 Jun 2007 10:29:23 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: <3AFlVjIU+Z5K@eisner.encompasserve.org> In article , Ron Johnson writes: > > Any dedicated newsreader worth it's salt has an "Ignore thread" feature. But any poster worth listening to will stick to the newsgroup topic. ------------------------------ Date: Thu, 07 Jun 2007 10:39:32 -0500 From: Ron Johnson Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: On 06/07/07 08:43, Paul Anderson wrote: > In article , > Ron Johnson wrote: > >> Any dedicated newsreader worth it's salt has an "Ignore thread" >> feature. > > I heard that doing that contributes to global warming. Now *that* was funny. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 07 Jun 2007 11:06:47 -0500 From: Ron Johnson Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: On 06/07/07 10:29, Larry Kilgallen wrote: > In article , Ron Johnson writes: >> Any dedicated newsreader worth it's salt has an "Ignore thread" feature. > > But any poster worth listening to will stick to the newsgroup topic. Did you misunderstand "newsreader" to mean "person who reads news"? -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 07 Jun 2007 18:17:30 +0200 From: Wilm Boerhout Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: <46682f9a$0$25478$ba620dc5@text.nova.planet.nl> on 7-6-2007 18:06 Ron Johnson wrote... [nip] > Did you misunderstand "newsreader" to mean "person who reads news"? > I, that is, me, as a person, have always had the "ignore thread" feature. Other newsreaders are slowly catching up... /Wilm ------------------------------ Date: Thu, 07 Jun 2007 10:43:17 -0700 From: Doug Phillips Subject: Re: HOWTO: Ignoring wildly OT threads Message-ID: <1181238197.366810.278840@p77g2000hsh.googlegroups.com> On Jun 7, 8:43 am, Paul Anderson wrote: > In article , > Ron Johnson wrote: > > > Any dedicated newsreader worth it's salt has an "Ignore thread" > > feature. > > I heard that doing that contributes to global warming. > Only if we don't observe the thread. Weisenheimer's theory of thread entanglement says that an off-topic thread in one newsgroup will spawn an OT discussion in another group and lead to global warming if the OT thread is not observed. Thread noodling by the Flying Spaghetti Monster can cause the entanglement to appear random. So says the great Weisenheimer. ;-) ------------------------------ Date: Thu, 07 Jun 2007 02:39:03 -0400 From: JF Mezei Subject: Re: Impact of an abnormally high RECNXINTERVAL Message-ID: Michael Moroney wrote: > I'll tell you what you'll see. The cluster will quickly "hang" and remain > hung until you either reconnect the cable or RECNXINTERVAL times out. > It'll look like a quorum hang, but it's not. Would a process whose resources are all local have the ability to continue to work, or would it also freeze because the operating system itself would hang ? (for instance, would having an unusually long RECNXinterval be a way to have a monitoring process continue after what would have been a loss of quorum so it can then take action (such as send a pager message) ? If a process has already assigned all channels prior to the loss of etehrenet, wouldn't it be able to send data out a serial port (to send a pager command for instance) ? ------------------------------ Date: Thu, 7 Jun 2007 16:41:34 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: Impact of an abnormally high RECNXINTERVAL Message-ID: JF Mezei writes: >Michael Moroney wrote: >> I'll tell you what you'll see. The cluster will quickly "hang" and remain >> hung until you either reconnect the cable or RECNXINTERVAL times out. >> It'll look like a quorum hang, but it's not. >Would a process whose resources are all local have the ability to >continue to work, or would it also freeze because the operating system >itself would hang ? I *believe* that as long as the process doesn't touch any locks that may depend on the disconnected node, it can continue to run. I don't know for sure since the situation I used large RECNXINTERVAL values (debugging kernel mode cluster code with XDELTA or the HALT button) it was on test clusters which I didn't care what processes on other node(s) were doing. >(for instance, would having an unusually long RECNXinterval be a way to >have a monitoring process continue after what would have been a loss of >quorum so it can then take action (such as send a pager message) ? >If a process has already assigned all channels prior to the loss of >etehrenet, wouldn't it be able to send data out a serial port (to send a >pager command for instance) ? I believe that would work. (your next challenge: detect such a situation without hanging the process) There *may* be a way for a process to be allowed to run without quorum. I've seen a process flag with quorum in its name, that all processes normally have, that seems to indicate the process can only run when the cluster has quorum. ------------------------------ Date: Thu, 07 Jun 2007 11:35:39 -0600 From: Keith Parris Subject: Re: Impact of an abnormally high RECNXINTERVAL Message-ID: Michael Moroney wrote: > JF Mezei writes: >>What happens when , with recnxinterval of 1 hour, I pull the ethernet >>cable out. Would all nodes continue to purr happily for one hour >>(instead of freezing due to loss of quorum) ? In such a case, I assume >>individual processes would go into LEF the second they do an IO on a >>remote disk ? > > Your cluster would likely "freeze" very quickly because of all the lock > requests not being granted (because of being in an incompatible mode > on the disconnected system, or the system mastering the particular lock) Or if you do a new lock request on a resource tree on which the node you're on doesn't already have any locks, and the lock directory lookup just happens to land on the disconnected system. JF Mezei writes: >Would a process whose resources are all local have the ability to >continue to work, or would it also freeze because the operating system >itself would hang ? It should be able to continue to run. Michael Moroney wrote: >There *may* be a way for a process to be allowed to run without quorum. >I've seen a process flag with quorum in its name, that all processes >normally have, that seems to indicate the process can only run when the >cluster has quorum. Right. In theory you could call the system service to remove your process's requirement for the QUORUM capability bit to be present on a CPU in order for your process to run, allowing your process to be scheduled and run on a CPU even during quorum loss, but then you'd have to be sure you didn't access any resources protected by the quorum scheme. This might be a way for a real-time process monitoring and controlling some sort of external equipment to continue to run despite a temporary quorum loss. But it wouldn't be able to do any disk I/Os (it would probably have to communicate via memory with another process to do any I/Os). And this wouldn't work if you depend on the quorum scheme to prevent two copies of any such process from running at once and trying to control the external equipment at the same time. ------------------------------ Date: Thu, 07 Jun 2007 08:21:57 -0700 From: etmsreec@yahoo.co.uk Subject: Re: Not on latest Roadmap: OpenVMS VAX Version 8.x "under investigation" Message-ID: <1181229717.420539.277350@k79g2000hse.googlegroups.com> Were the two nodes actually doing NFS writes at the time of the network outage? If they were, that could go some way to explaining the problem... Steve On 6 Jun, 21:41, Ron Johnson wrote: > On 06/06/07 15:14, Keith Parris wrote: > > > > > > > Ron Johnson wrote: > > {an excellent and informative description of why he prefers Linux to > > *BSD, followed by:} > >> just the other day at work, THREE Alphas (2 clustered and > >> one stand-alone, all running v8.2) crashed during a network hiccup > >> while one of the clustered boxes and the stand-alone box were > >> writing to an NFS drive. The third box (the other cluster member) > >> crashed *hard* when it lost cluster connectivity. > > > To allow cluster members to ride through a temporary LAN outage (where > > the LAN is used as the cluster interconnect), folks normally raise the > > SYSGEN parameter RECNXINTERVAL to a value just over the maximum number > > of seconds in duration of an outage they need the cluster to ride through. > > > I wouldn't expect a standalone VMS node to crash on a LAN outage. Did > > you get crash dumps? > > On one of the nodes, but not the other. Our SysAdmin sent it off to HP. > > > And I wouldn't expect a cluster member to crash > > when it _lost_ cluster connectivity > > Neither would I. :( > > > -- it should suspend its operations > > due to quorum loss until connectivity is restored, and then it might > > crash if it found it had been removed from the cluster by the other > > members in the intervening time. > > -- > Ron Johnson, Jr. > Jefferson LA USA > > Give a man a fish, and he eats for a day. > Hit him with a fish, and he goes away for good!- Hide quoted text - > > - Show quoted text - ------------------------------ Date: Thu, 07 Jun 2007 10:43:03 -0500 From: Ron Johnson Subject: Re: Not on latest Roadmap: OpenVMS VAX Version 8.x "under investigation" Message-ID: One was writing; it crashed first. The other node crashed because of lost connectivity with the first node. I'm not too up on the details, though. On 06/07/07 10:21, etmsreec@yahoo.co.uk wrote: > Were the two nodes actually doing NFS writes at the time of the > network outage? If they were, that could go some way to explaining > the problem... > > Steve > > On 6 Jun, 21:41, Ron Johnson wrote: >> On 06/06/07 15:14, Keith Parris wrote: >> >> >> >> >> >>> Ron Johnson wrote: >>> {an excellent and informative description of why he prefers Linux to >>> *BSD, followed by:} >>>> just the other day at work, THREE Alphas (2 clustered and >>>> one stand-alone, all running v8.2) crashed during a network hiccup >>>> while one of the clustered boxes and the stand-alone box were >>>> writing to an NFS drive. The third box (the other cluster member) >>>> crashed *hard* when it lost cluster connectivity. >>> To allow cluster members to ride through a temporary LAN outage (where >>> the LAN is used as the cluster interconnect), folks normally raise the >>> SYSGEN parameter RECNXINTERVAL to a value just over the maximum number >>> of seconds in duration of an outage they need the cluster to ride through. >>> I wouldn't expect a standalone VMS node to crash on a LAN outage. Did >>> you get crash dumps? >> On one of the nodes, but not the other. Our SysAdmin sent it off to HP. >> >>> And I wouldn't expect a cluster member to crash >>> when it _lost_ cluster connectivity >> Neither would I. :( >> >>> -- it should suspend its operations >>> due to quorum loss until connectivity is restored, and then it might >>> crash if it found it had been removed from the cluster by the other >>> members in the intervening time. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 07 Jun 2007 02:44:57 -0400 From: JF Mezei Subject: Re: porting ignorance (was Re: Story Time) Message-ID: Craig A. Berry wrote: > statistics and a dictionary. Then there are things like Apache and > Mozilla that took a very substantial effort from OVMS Engineering. I suspect that gathering all the pieces is the most time consuming piece (as opposed to actually modifying code). For instance, just compiling Mozilla requires you have a half dozen anciliary packages such as perl etc etc (and you probably also need a fair amount of utilities to unpack from whatever source packaging is used (unless you are lucky enough to get it in .zip form). And since Mozilla needs GTK, you also need to build GTK, and to do so, they probably eithe had to hack GTK or X-windows because the X-window on VMS isn't exactly current. ------------------------------ Date: 7 Jun 2007 09:53:40 +0200 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOeGER) Subject: Re: porting ignorance (was Re: Story Time) Message-ID: <4667d5a4$1@news.langstoeger.at> In article , JF Mezei writes: >Craig A. Berry wrote: >> statistics and a dictionary. Then there are things like Apache and >> Mozilla that took a very substantial effort from OVMS Engineering. > >I suspect that gathering all the pieces is the most time consuming piece >(as opposed to actually modifying code). For instance, just compiling >Mozilla requires you have a half dozen anciliary packages such as perl >etc etc (and you probably also need a fair amount of utilities to unpack >from whatever source packaging is used (unless you are lucky enough to >get it in .zip form). > >And since Mozilla needs GTK, you also need to build GTK, and to do so, >they probably eithe had to hack GTK or X-windows because the X-window on > VMS isn't exactly current. And this (MOZILLA) was all done by one man only which is unfortunately no longer with HPQ/this duty... -- Peter "EPLAN" LANGSTOEGER Network and OpenVMS system specialist E-mail peter@langstoeger.at A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Thu, 7 Jun 2007 08:15:02 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: porting ignorance (was Re: Story Time) Message-ID: In article , "Craig A. Berry" writes: > To whomever isn't doing something non-trivial in his or her spare time, > I pity you. And anyone who thinks that all porting work that is of > value is non-trivial (or is done entirely by volunteers in their spare > time) is not well-informed about either porting or about open source > software. > These matters are well known and frequently discussed by people who > actually work with open source software instead of talking about how > difficult it is to port to VMS. > I haven't followed this whole thread (I'm too busy porting open source > software to read everything here), but it certainly appears that the > main thing thing being ported and re-ported here is ignorance. I'm reminded of a quote by Dave Jones (author of the "OSU" HTTP server): BTW, the source code to perl gives me a headache, I haven't seen anything so convoluted since I did the Maple ports for Waterloo. The comments are either quotes from 'Lord of the Rings' or self-congratulatory remarks of how efficient or clever the next block of code is. ------------------------------ Date: Thu, 07 Jun 2007 06:17:59 -0500 From: Ron Johnson Subject: Re: porting ignorance (was Re: Story Time) Message-ID: On 06/07/07 03:15, Phillip Helbig---remove CLOTHES to reply wrote: [snip] > > I'm reminded of a quote by Dave Jones (author of the "OSU" HTTP server): > > BTW, the source code to perl gives me a headache, I haven't seen > anything so convoluted since I did the Maple ports for Waterloo. The > comments are either quotes from 'Lord of the Rings' or > self-congratulatory remarks of how efficient or clever the next block > of code is. Which is why they are writing Perl6 from scratch (and it's taking a *long* time). -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 07 Jun 2007 13:47:02 +0200 From: "P. Sture" Subject: Re: porting ignorance (was Re: Story Time) Message-ID: In article , JF Mezei wrote: > Craig A. Berry wrote: > > statistics and a dictionary. Then there are things like Apache and > > Mozilla that took a very substantial effort from OVMS Engineering. > > > I suspect that gathering all the pieces is the most time consuming piece > (as opposed to actually modifying code). For instance, just compiling > Mozilla requires you have a half dozen anciliary packages such as perl > etc etc (and you probably also need a fair amount of utilities to unpack > from whatever source packaging is used (unless you are lucky enough to > get it in .zip form). I've often found it convenient to unpack distributions on OS X, zip the results there, then unzip on VMS. A couple of points with that though: In OS X you can often simply double click on the compressed file, and Finder will unpack it for you. You can simply use Finder's Archive function to create a zip file which VMS UNZIP can (usually) read. However, with a typical .tar.gz file., that method will leave behind the intermediate file, where the equivalent won't. A real problem here with certain archives is that OS X Finder can "get carried away" and go too far with unpacking, resulting in an incomplete directory tree (imagine BACKUP/RESTORE overwriting a complete directory if it already exists!!!). IOW, it's best to learn the correct syntax for tar and friends and use that at the CLI instead of relying on Finder's behaviour. -- Paul Sture ------------------------------ Date: Thu, 07 Jun 2007 09:33:19 -0700 From: davidc@montagar.com Subject: Re: porting ignorance (was Re: Story Time) Message-ID: <1181233999.420301.304740@q75g2000hsh.googlegroups.com> On Jun 6, 10:48 pm, "Craig A. Berry" wrote: > In article <5cl99aF307t9...@mid.individual.net>, > b...@cs.uofs.edu (Bill Gunshannon) wrote: > > > Someone else hinted that porting > > Unix/Linux Open Source Software was somehow trivial and could be > > done by the people here in their spare time. > > To whomever isn't doing something non-trivial in his or her spare time, > I pity you. And anyone who thinks that all porting work that is of > value is non-trivial (or is done entirely by volunteers in their spare > time) is not well-informed about either porting or about open source > software. Never said it was trivial. But I do believe that in many/most cases it is possible and worthwhile. I agree that sometimes very complicated things work trivially, while some simple things end up being a pain. In some cases, it might be worthwhile to just rewrite it for VMS. As noted by others, sometimes it isn't the app that's the problem, but the dependencies on a bunch of other smaller packages. You can't port the app without also porting the dependencies. This is where I think we can provide benefit, by helping with some of these dependencies and creating a growing library of these that continue to provide base support, and feeding back updates to the maintainers to provide for future VMS support. Things like this have been done for Samba and other projects, where you can download from the main site, and theres a BUILD.COM or DESCRIP.MMS right there in the distribution. I'm sure there would be little argument from many of these projects if you asked them to include a functional DESCRIP.MMS and VMS config.h file as part of the codebase. Since many of these libraries are commonly used in a wide variety of apps/libraries, a lot of good is possible from it. Any not just for the sake of porting OSS apps. Many of these libraries are fairly native on other platforms, and making them "native" on VMS further eases software portability to VMS, too. ------------------------------ Date: Thu, 7 Jun 2007 08:10:33 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: SECURITY.AUDIT$JOURNAL and ERRLOG.SYS Message-ID: In article <5cpb5dF31vmcgU1@mid.individual.net>, Ken Fairfield writes: > Since I was mounting the Alpha system disk as the cluster-common > disk on VAXes, I needed to both redefine the SYS$SYSROOT search-list > _and_ mount the disk as early as possible (basically at the same time). > That put some "odd" code into SYLOGICALS to loop with F$GETDVI > waiting for the device to EXIST, then to be AVL (or AVAIL?), and > then mount it. This all from memory of 7 years ago or so... > Although I did post more than once to c.o.v about my technique, > so the archives would be more accurate than this description. Right. I think I picked up some of that code and adapted it to my needs. ------------------------------ Date: 7 Jun 2007 13:15:32 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Story Time Message-ID: <5cqenkF2vlshjU1@mid.individual.net> In article , Ron Johnson writes: > On 06/06/07 17:50, Bill Gunshannon wrote: >> In article <6qF9i.92829$vE1.77432@newsfe24.lga>, >> Ron Johnson writes: >>> On 06/06/07 13:09, Bill Gunshannon wrote: >>>> In article , >>>> Ron Johnson writes: >>>>> On 06/06/07 10:30, Bill Gunshannon wrote: >>>>>> In article , >>>>>> Ron Johnson writes: >>>>>>> On 06/06/07 07:58, Bill Gunshannon wrote: >>>>>>>> In article , >>>>>>>> Ron Johnson writes: >>>>>>>>> On 06/05/07 13:52, Bill Gunshannon wrote: >>>>>>>>> [snip] >>>>>>>>>> All of which applies to BSD equally except that BSD had several years >>>>>>>>>> headstart (including the development that continued despite the AT&T >>>>>>>>>> lawsuit which everyone involved in the technical side of the game knew >>>>>>>>>> was never going to go anywhere). The only thing Linux has that BSD >>>>>>>>>> does not is marketing. And look at the difference in awareness and >>>>>>>>>> interest. BSD's license is much more business friendly than the GPV. >>>>>>>>> Two words: Unix Wars. >>>>>>>> People keep mentioning that but unless your a techie it really means >>>>>>>> nothing. >>>>>>> I *am* a techie, and I *do* remember when all the various vendors >>>>>>> took BSD or licensed SVRx and "compatible" C & Unix became a >>>>>>> mismash, and then there was the OSF, Unix International, etc, none >>>>>>> of which really unified Unix. >>>>>> Ancient history. And irrelevant to the discussion of current BSD vs. >>>>>> Linux. >>>>> I think that it *is* relevant because it was the "business friendly" >>>>> BSD license (which allows each company to keep it's own changes) >>>>> that caused the Unix Wars in the first place. >>>> What you call the Unix Wars is nothing different than the battling >>>> between all proprietary OSes. It's called business and yes, it is >>>> war. And while many of those Unixes are still around and kicking >>>> today they don't enter into the debate because we were comparing >>>> Free Unixes. A CIO who is going to buy AIX is going to buy AIX and >>>> the superiority of BSD over Linux doesn't enter into it. But when >>>> one is going to base their strategy on a Free Unix then one has to >>>> ask why one over the other and more importantly, why did the inferior >>>> product win the market? The only good thing that Linux has over BSD >>>> is hype and that is totally the result of strong marketing. >>>> >>>>> The (smart) big vendors remember these things. >>>>> >>>>> Counterintuitively, the "viral" GPL (which says, in essence, "freely >>>>> you get other people's work, freely you must share your own work") >>>> Is that like forced volunteerism that is all the rage today? I can't >>>> be coerced into giving something "freely". >>> Coerced? No. It's called rule of law. >> >> It has nothing to do with law other than relying on it to support them. >> When you put strings on the use of something it isn't free. You said, >> "If I do a then I must do B". That takes away my freedom to not do B. >> That's coercion. > > Like HP coerces you not to give VMS licenses away for free? I guess I'm just dense, but I don't get this at all. Who other than HP (and its authorized agents) gives out VMS licenses at all? And since when is VMS "free software"? The Hobbyist Program not withstanding, VMS is pure proprietary commercial software. (Althougn lately based on comments I have seen I wonder how many people might be running commercial operations using Hobbyist PAKs.) > >>> To use any code that is not public domain, you must agree to the >>> license. In this case, the General Public License says that "you" >>> can freely use "his" code, but that if you distribute the subsequent >>> binaries to anyone else, you must share your changes with "them". >>> >>> Think of it as payment for "his" code. >> >> Payment? I thought this was "free" software. That is what they >> like to claim. > > Shows how little you know of the GPL. Are you saying they don't claim GPLed software is "free"? > > The phrase usually used it "free as in speech, not (necessarily) > free as in beer". In fact, the GPL encourages you to charge money > for shipping your code to your distributees. "free" is "free". Why hide behind fancy rhetoric unless you have something to hide? Is it "free" software or not? Code released under the BSD license is "free", the only strings attached relate to the text of the copyright message and in no way restrict what you may or may not do with the code. The only thing more free is public domain. > >> Fact is, it is encumbered just like the old AT&T >> code was. The only difference is AT&T openly claimed theirs was >> encumbered and GNU claims theirs is free. >> >>>>> ensures that the Unix Wars can not happen again. >>>> Not hardly. There is at least one distro that has two versions. One >>>> they give away for free and the other you have pay for. And they openly >>>> admit the two are different. And the one you pay for has additional >>>> features not in the free one. >>> RH and SuSE do that. >> >> You seem to take that rather calmly for someone who seems to think the >> BPL is a good thing, but isn't that contrary to the GPL? > > BPL?? Sorry, mistype. But I would have hoped it was obvious considering the proximity of the "G" and "B" keys and what were discussing. > > None of these definitions make sense in the current context. > > Broadband over Power Lines > BPL group, an Indian electronics conglomerate > Boston Public Library > Brooklyn Public Library > Berkeley Public Library > Borland Package Library > BPL (complexity), the complexity class of problems solvable > with bounded-error randomized algorithms in logarithmic space > Brass Pounders League, a term for Morse code operators that > comes from the use of brass Telegraph keys > Below the poverty line, a term for levels of income below the > Poverty threshold > Burnaby Public Library, a public library located in Burnaby, > British Columbia > Basic perturbation lemma, used in computer science and algebra > Brisbane Premier League, a football (soccer) competition in > Brisbane, Australia. > > >>>> And there are other GNU Programs that >>>> while living up to the letter of the agreement do not live up to the >>>> spirit in that they have made the source code they provide useless >>>> without paying them for the tools to work with it. > >> >>> Tivo? >> >> Don't know about Tivo as not being a television addict I have never >> used one. I was thinking of something totally software. > > It runs on Linux (and that code is available to customers) but you > can't modify the code, because Tivo will only run if the hash key of > the compiled binary matches a pre-calculated number. So, it is like I said, the GPL didn't protect the "freedom" of this code. Just one more to add to my long (and growing) list. > >>> The RH "commercial" version does have some closed-source features, >> >> And this doesn't violate the GPL? Strange! > > Absolutely not. > > The GPL'ed code is available, and the closed-source code isn't. I thought modification to GPLed code had to be GPLed? > > Just like running Oracle on Linux. We are not talking about none GPLed applications here. There are difference ins the kernel between the two versions. The kernel is GPLed and any modification are supposed to be GPLed as well. > >>> but CentOS makes a (rather popular) fully functional distro using >>> the same sources that RH Advanced/Enterprise Server uses. >> >> Oh great, yet another Linux distro. What was it you said about Unix Wars? > > And it's *fully* source-and-binary compatible with Red Hat. It's still yet another fork in the trail. Something that many call a bad thing for BSD. So, how long before it stops being compatible because they decide they do not agree with the direction RH is going? Or will the GPL prevent that? :-) > >>>>> Thus, it's sort of a neutral platform where binary compatibility is >>>>> almost guaranteed. >>>> Your joking, right? Debian won't run RedHat. RedHat won't run Slackware, >>>> etc. >>> There *is* a high likelihood, if all the distro versions are of the >>> same vintage. >> >> A high likelihood is not much use if it turns out you have the wrong >> version. I have Knoppix on a mach8ine at home for testing things and >> I find very little that will actually work with it. > > Running off a CD/DVD? No, installed on a hard disk from a CD. (One thing that Linux distros seem to do well is install. Although the latest CD from Sun was also impressive!) > >> And then, I have >> my Nokia 770 which is Linux based and is very selective about what it >> will run. > > A 10 second search brought me here: > http://en.wikipedia.org/wiki/Nokia_770 > > Which brought me here: > http://maemo.org/community/wiki/ApplicationCatalog > > And that lists a *lot* of applications. Many of which won't work depending on which version of the kernel you are running. Many more that have dependancies that are not avaialable or only available for money. Or, in many cases just flat out don't work. Most break with every update to the machine. Nice toy, but still not practical for reliable use. Don't get me wrong, I like mine and now that there is a version of Maemo Mapper that works with mine again I use it frequently instead of actually buying a Garmin. :-) > >>> [snip] >>>>>>> Overall, though, you won't hear me complaining because you run FreeBSD. >>>>>> Well, you certainly won't hear me complain. We once tried to use Linux >>>>>> to do the job because people wanted the more popular option. It took >>>>>> less than one semester to have all of them learn what a mistake Linux >>>>>> really was. We have never looked back and Linux will never have a >>>>>> place in our server farm as long as I am the Admin here. >>>>> I'd *REALLY* like to hear what the problem with Linux was. > >>> >>>> Are you really serious? Broken LPD. >>> lpd? Who in Linux uses lpd? >> >> Anybody who has to support a heterogenous network. Even MS got it right >> but not Linux. > > OK, point taken. > >>>> Badly broken NFS. >>> Yes, I've heard it's not the most efficient. >> >> Efficiency is one matter, broken is another. (Hint: set up a test >> network. NFS mount a directory. Boot the NFS Server. All of the >> clients will need to be re-booted before they can access the file- >> system again. NFS is supposed to be stateless!!) > > Well then Linux is in good company: an NFS disconnect crashed an > AlphaVMS 8.2 cluster last week. Well, I would need more info to know who was actually at fault on this one, but just to keep the record straight, as much as I like VMS I have never bought into this "uncrashable/unhackable/guaranteed 24-7" nonsense. It's an OS, not god. I have seen VMS crash. I have seen "trojans" run on VMS (when I first got here it was one of the most common things run by students in the terminal rooms!) I have watched the annoucements on login that the local system would be down on such-and-such a day for maintenance. > > But seriously: > Kernel-mode or user-mode NFS? > v2 or v3 or v4? Sorry, I don't research it. I just watched it not work and found a replacement. Not that it makes a difference. NFS is stateless. The behavior on Linux was something I have never seen on any other version of NFS and I have used NFS on: SunOS, Solaris, lots of BSD's, Irix, HPUX, Ultrix-32, MSDOS, Xenix, QNX and I am sure others that I have long ago forgotten. So, when the NFS server crashes how do you tell all your users that you have to reboot the clients and they are going to loose all their work up to that point because they can't save their file? > > Or it might have been a NIC problem. Look at this posting, for example. > https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.17/+bug/71212/comments/16 No, not a NIC problem. It is a state problem in something designed to be stateless. I can duplicate it at will and it behaves the same no matter what particular Linux Distro is used. Plain and simple, they got it wrong and probably have no intention of fixing it because they think it is everyone else that is wrong. > >>>> Extremely >>>> inefficient IP stack because of NIH Syndrom. >>> None of the benchmarks I've seen show that it's "bad". Maybe the >>> particular driver for the NICs you were using... >> >> Must be using the wrong benchmarks. Every one I ever ran showed Linux >> way behind. > > Got any urls? No, really. The kernel team *always* wants to fix > problems. Then let them do their own work. I have no interest in fixing Linux. I prefer to use other options that just work. > >>> [snip] >>>>>> But, to reiterate the original theme, if marketing can make such a >>>>>> success out of a piece of crap like Linux, just think what it could, >>>>>> no would do for a gem like VMS. >>>>> If techies (not sheeple) really thought that Linux was a steaming >>>>> pile of dung, it wouldn't have lasted this long. >>>> What techies? All those prepubescent teens without girlfriends who >>>> put up a website in their parents basement during the dot.com boom >>>> who now skew the unemployment numbers by claiming to be "un-employed >>>> IT professionals" today? >>> Great ad hominem, but, ummm, no. >> >> Is an "ad hominem" still an "ad hominem" when it's true? We have had >> the chance to see lots of these geniuses show up here. Freshman who >> want credit for courses without taking them becuas ethey have been >> running "businesses" since they were high school sophomores. They >> seldom do well and in a number of cases have either changed major to >> something other than CS/CIS or just plain dropped out. > > I can't comment, since I don't know any such people. Try going out on the Web and reading some resumes of "un-employed IT professionals". Some of them are funnier than rec.humor bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 7 Jun 2007 13:44:31 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Story Time Message-ID: <5cqgdvF31bpjmU1@mid.individual.net> In article <1181168029.866585.96120@o11g2000prd.googlegroups.com>, davidc@montagar.com writes: > On Jun 5, 1:52 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> All of which applies to BSD equally except that BSD had several years >> headstart (including the development that continued despite the AT&T >> lawsuit which everyone involved in the technical side of the game knew >> was never going to go anywhere). The only thing Linux has that BSD >> does not is marketing. > > False. The GPL attracted a lot of people to the Linux platform. The > BSD license allows a company to profit off the work of someone else > without compensation or credit. The GPL doesn't allow this. You > either get credited and/or compensated. No one is arguing this. the argument is which one is "free" software and which one is "encumbered" software. People are free to use any license they wish, but lying about it in orer to mislead the gullible is just plain wrong. Heck, if you want you can use a license that says people can only use your code naked if you want, just don't say it is "free". And using this "free speech"/"free beer" game is just trying to put up a smoke screen. Makes you think they are lawyers rather than computer geeks. > >> And look at the difference in awareness and >> interest. BSD's license is much more business friendly than the GPV. >> BSD is stabler, more secure, more efficient and has more stuff that >> was actually implemented correctly than Linux. And still businesses >> are flocking to Linux and ignoring BSD. And the answer is, marketing. >> Ask any CIO you know who is involved in one of these Linux migrations >> why Linux and not BSD. The most likely answer will be, "What's BSD?" >> Sure sounds like the same boat VMS is in to me. :-) > > That would be G. P. L., and yes the BSD license is more friendly to > the business, but to the individual programmers, the GPL can often be > more attractive, for reasons stated above. Word of mouth is what is > at work here. As developers become interested in the platform, they > play with it, books get written, the interest feeds upon itself. I > submit to you that Linux would have succeeded regardless of IBM's > marketing dollars. I wasn't talking about IBM. They came into the game late, long after Linux was popular. SWhen I said marketing I meant thngs like "The Linux Journal" and all the effort that was made by smalltime players getting articles about Linux into trade journals. And that had a snowball effect. Every article made the trade rags want to find more articles. If BSD had made the same effort at marketing I seriously expect that Linux would have died very early in it's life. > > During the developement of Linux, Bill Gates spent more money wiping > his butt with $20 bills than Linux had in marketing. But that's the cool thing about marketing (and the thing that HP just can't seem to grasp concering VMS). It doesn't have to cost money. It can be free (articles in trade rags) or even a money maker (The Linux Journal). It just takes a little interest in sowing the seeds and soon it can take on a life of its own. Linux today is cool and big business. An article written for Computer World about Linux by anyone with even a minor reputation as an IT Journalist is going to get published. > Yet, you didn't > see developers give up, wail about lack of marketing dollars, and > bitch and moan about it in forums. If what you believe were true, > Linux NEVER would have succeeded. But it did. If all that work > hadn't have been done FIRST, there wouldn't have been anything to > market at all. Linux was already being hyped as the best thing since sliced bread long before it was viable as a commercial OS and BSD was already way ahead of it in stability, features and available applicatations. Technical merit has never been the deciding factor the success of this kind of product. BSD already had a network of developers world- wide before Linus wrote his first line of code. > > Why it succeeded was because of the community. see the above comment. I'll say it again. BSD already had a network of developers worldwide before Linus wrote his first line of code. > I believe that is > going to be what makes VMS stronger, too. A good solid community of > developers and users working toward the goal regardless of what HP > does. If there aren't tools and apps ported to OpenVMS, then OpenVMS > is an O/S with no purpose - and no amount of marketnig dollars can fix > that. The product, and assoicated side products, must be there > first. But if you can't lead or follow, at least don't start by > assuming there's no point in trying. Nice thought, but not reality. The fate of VMS is toatlly in the hands of HP. The community can do anything it wants to put apps on VMS but unless HP is willing to sell it to potential customers (and such customers exist) VMS will wither on the vine, except maybe in the basements of hobbyists. > > That's why HP has been doing programs like the HP Integrity Developers > Workshops (which, I think is a marketing effort costing in the > hundreds of thousands of dollars), That's not marketing, that's preaching to the choir. > free downloads of software for > DSPP, free Integrity SDK kits shipped to you simply for the asking, > and more. But only the VMS faithful know about any of this. > > But I'm sorry Bill. That probably isn't enough for you. Ummm.... I am not the only one (here or elsewhere) that thinks more is needed. I have watched VMS die here. What VMS is still here is hidden behind a smoke screen of Java and Windows. Where everyone used to know what was driving the University's IT today it's all Windows. Something about a light under a basket comes to mind. How can oyu sell something if you don't tell potential customers it's avaialble? You can bet that Sun and IBM are telling people. We dumped Sun over 15 years ago and yet I still get periodic calls from their sales people trying to see if there is any interest. All I have gotten from HP in the past 4-5 years a solid demonstration that they don't understand my business which is hardly likely to get them any. And I am one of the people who fought to keep them here as long as possible!! > >> I am not spinning anything. Name the applications that people are most >> likely to want. Then look at what they contain. Actually, the obsolete >> version of X11 that is available for VMS is probably the bigger show >> stopper as desktop apps are what sells computers today. But there are >> still alot of cute little things with fork() in them. And that list >> was not meant to be exclusive. It just pointed out the two most common >> shortcomings frequently mentioned here everytime this subjecy comes up. > > Yes you are. You are already predicting failure of porting software > even before anyone STARTS. You use two issues and predict 100% > failure right out of the gate. I guess you aren't capable of working > around fork(), but please don't convince others they can't before they > start. Most of the time fork() is called, it's to detach a process > anyway (fork followed by exec) and can be replaced with a call to LIB > $SPAWN or SYS$CREPRC. As for X11, you are seriously out of touch > since rarely do people program X11 calls. The only ones that do are > those developing window managers or maintaining the GUI toolkits like > GTK or wxWidgets. The rest are using those toolkits like GTK+ or > wxWidgets. OK, you win. I have better things to do with my time. Let me know when you have OpenOffice running on VMS. (that's the kind of apps that real users want, not PHP.) bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Thu, 07 Jun 2007 07:40:32 -0700 From: Andrew Subject: Re: Story Time Message-ID: <1181227232.012679.272560@o5g2000hsb.googlegroups.com> On 7 Jun, 00:33, "P. Sture" wrote: > In article , > Ron Johnson wrote: > > > VMS is secure from the get-go, but Linux not so much. Is SELinux > > enabled? If so, then what profiles are enabled/created? (SELinux > > was designed to bring mainframe/VMS-style security to Linux.) > > is SELinux widely available, and if so, widely adopted? > It is widely available but it tends to get turned off because many admins struggle to get it to work. RedHat have started a campaign to get admins to use it rather than turn it off which is getting a mixed reception. There is currently a discussion about it on Slashdot. http://it.slashdot.org/article.pl?sid=07/06/06/007218 Regards Andrew Harrison ------------------------------ Date: Thu, 07 Jun 2007 08:18:55 -0700 From: davidc@montagar.com Subject: Re: Story Time Message-ID: <1181229535.087418.297450@p47g2000hsd.googlegroups.com> On Jun 7, 8:44 am, b...@cs.uofs.edu (Bill Gunshannon) wrote: > Nice thought, but not reality. The fate of VMS is toatlly in the > hands of HP. The community can do anything it wants to put apps > on VMS but unless HP is willing to sell it to potential customers > (and such customers exist) VMS will wither on the vine, except maybe > in the basements of hobbyists. Read that carefully - VMS, by itself, is simply a platform. If HP sells you the platform, what are you going to do with it? You need support from a developer community so you can sell a solution that does something. There's an active developer community, and tons of code, out there able to be capitalized on. I think we, as a community, can help take advantage of that. Although you dismiss them, the ports of Apache, MySQL, PHP are an example of some of the enabling technologies where OpenVMS is actively competing in the data center against Sun, MS, IBM, and others. There's an incredible amount of PHP code available for doing cool things, and having it backed by the extra security of OpenVMS is a value added. And commercially, are probably where most of the Linux installations are situated, not the desktop. > > That's why HP has been doing programs like the HP Integrity Developers > > Workshops (which, I think is a marketing effort costing in the > > hundreds of thousands of dollars), > > That's not marketing, that's preaching to the choir. Okay, so you think they shouldn't do this program? I think it's a benefit since it promotes continued development, and the free hardware doesn't hurt. > > > free downloads of software for > > DSPP, free Integrity SDK kits shipped to you simply for the asking, > > and more. > > But only the VMS faithful know about any of this. Actually, I've told many people about it. In fact, that's actually one of the "intentions" of the Hobbyist Program, is to encourage use of OpenVMS, and even software development on OpenVMS. Of course, if you develop a commercial product for OpenVMS, you'll need to sign up for DSPP or something to get non-Hobbyist licenses before you sell the first copy, but DSPP is pretty cheap, too. > > But I'm sorry Bill. That probably isn't enough for you. > > Ummm.... I am not the only one (here or elsewhere) that thinks more > is needed. Not saying it isn't. It's just I think we can be part of the solution rather than part of the problem. Actually do something rather than complain about it. > OK, you win. I have better things to do with my time. Let me know > when you have OpenOffice running on VMS. (that's the kind of apps > that real users want, not PHP.) No, Bill. You win. Even when OpenOffice is running on OpenVMS (and as mentioned, that effort is in progress), you'll still find something else you complain about. Part of the issue is that even the Linux camp is wondering if Linux is ready for the Desktop against Windows. So even in the Linux world, OpenOffice isn't the yardstick of success, whereas for some reason you've chosen it as yours. In fact, I'd almost guess that you've selected it explicitly as the having the greatest chance of failure, thus affirming your claim of hopelessness. ------------------------------ Date: Thu, 07 Jun 2007 11:45:26 -0600 From: Keith Parris Subject: Re: Thin wire Coax Ethernet on Alpha Message-ID: Jim Mehlhop wrote: > Anyone know of a particular Ethernet board with Thin Wire Coax (10MB) > connectorr that will run on an Alpha?? I believe the DE425, DE435, and DE450 PCI cards would qualify. Check the Supported Options list for the particular Alphaserver model on the HP website. ------------------------------ Date: Thu, 7 Jun 2007 10:06:14 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: using RBLs interactively Message-ID: I have my TCPIP stuff set up to use ZEN.SPAMHOUSE.ORG as an RBL and it seems to work. However, when using the following procedure to query directly, I ALWAYS get 127.0.0.2 as a response. On the other hand, if I use the query form at spamhouse.org, I get the expected result. $ ADDRESS = P1 $ A = F$ELEMENT(0,".",ADDRESS) $ B = F$ELEMENT(1,".",ADDRESS) $ C = F$ELEMENT(2,".",ADDRESS) $ D = F$ELEMENT(3,".",ADDRESS) $ TCPIP SH HOST 'D'.'C'.'B'.'A'.'P2' ! adding a final "." makes no difference $ EXIT ------------------------------ Date: Thu, 07 Jun 2007 12:50:24 GMT From: "John E. Malmberg" Subject: Re: using RBLs interactively Message-ID: Phillip Helbig---remove CLOTHES to reply wrote: > I have my TCPIP stuff set up to use ZEN.SPAMHOUSE.ORG as an RBL and it > seems to work. However, when using the following procedure to query > directly, I ALWAYS get 127.0.0.2 as a response. On the other hand, if I > use the query form at spamhouse.org, I get the expected result. > > $ ADDRESS = P1 > $ A = F$ELEMENT(0,".",ADDRESS) > $ B = F$ELEMENT(1,".",ADDRESS) > $ C = F$ELEMENT(2,".",ADDRESS) > $ D = F$ELEMENT(3,".",ADDRESS) > $ TCPIP SH HOST 'D'.'C'.'B'.'A'.'P2' ! adding a final "." makes no difference > $ EXIT How does it work with zen.spamhaus.org ? -John wb8tyw@qsl.network Personal Opinion Only ------------------------------ Date: Thu, 07 Jun 2007 13:08:14 +0200 From: "P. Sture" Subject: Re: VMS Update Monday June 4th Message-ID: In article <1181186913.298750.110010@p77g2000hsh.googlegroups.com>, Sue wrote: > On Jun 6, 7:22 am, "P. Sture" wrote: > > In article <37b37$46662ce9$cef8887a$20...@TEKSAVVY.COM>, > > JF Mezei wrote: > > > > > Sue wrote: > > > > > > Did you even see the part about joining the DEC web site? > > > > > The one abouthttp://www.decedout.orgthat invites people to join it, > > > even if they are not ex-dec employees ? No, I didn't see it. :-) > > > > Membership appears to be for US and Canadian residents only. Is there > > any equivalent organisation for the rest of the world? > > > > -- > > Paul Sture > > Membership is worldwide I checked with them today. > Thanks Sue. -- Paul Sture ------------------------------ End of INFO-VAX 2007.310 ************************