1 INFO-VAX	Wed, 27 Dec 2006	Volume 2006 : Issue 713       Contents:0 Re: Cluster connection lost when one link fails?$ Re: DSSI and cluster quorum question$ Re: DSSI and cluster quorum question  MI5 Persecution: BBC h2g2 online* MI5 Persecution: their methods and tactics- Re: Need help with DHCP Client & Name servers A Re: Nonpaged pool expansion due to burst of OPCOM DECnet messages A Re: Nonpaged pool expansion due to burst of OPCOM DECnet messages   F ----------------------------------------------------------------------    Date: 27 Dec 2006 03:41:49 -0800/ From: "Volker Halle" <volker_halle@hotmail.com> 9 Subject: Re: Cluster connection lost when one link fails? B Message-ID: <1167219709.064075.117070@a3g2000cwd.googlegroups.com>   Malcolm,  E our IPMT has been closed by providing a new SYS$EWDRIVER_DE500BA.EXE, ? so the problem seems to have been in the underlying LAN driver.    Happy New Year,    Volker.    ------------------------------    Date: 27 Dec 2006 07:45:58 -0800/ From: "Volker Halle" <volker_halle@hotmail.com> - Subject: Re: DSSI and cluster quorum question C Message-ID: <1167234358.584304.224230@i12g2000cwa.googlegroups.com>    JF,   D DSSI is being used for cluster communications (same as good old CI).   Volker.    ------------------------------  + Date: Wed, 27 Dec 2006 17:03:15 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) - Subject: Re: DSSI and cluster quorum question ( Message-ID: <emu90j$ohj$1@pcls6.std.com>  0 JF Mezei <jfmezei.spamnot@vaxination.ca> writes:         >    +====DSSI===+  >[VELO]      [WHEEL]      [BIKE] >    +           +            +  >    ------------+------------+  >          Ethernet      >VELO has 2 votes  >WHEEL has 1 vote  >BIKE has 1 vote. 6 >No quorum disk.  So expected votes of 4, quorum of 3.    * >Ethernet to VELO is broken.  Quorum lost.  D >When ethernet is restored, only Bike commits hari-kiri and reboots.  I >How come WHEEL didn't also commit hari-kiri once ethernet was restored ?   M >Does the DSSI connection allow WHEEL to retain "sanity" and not be required  + >to reboot once quorum was re-established ?   N >I was under the impression that SCS did not use DSSI which served only disks.  J No, DSSI also carries cluster traffic.  You're probably thinking of shared SCSI being storage-only.  8 While in cluster transition, from the viewpoint of VELO,@ BIKE was kicked out of the cluster.  From the viewpoint of BIKE, VELO was kicked out.  E Once connectivity was restored, the inconsistent view of the cluster  D needed to be resolved.  Certainly it made more sense for a node with@ one vote to reboot than the node with two votes, and that's what	 happened.   C If there was no DSSI bus, you'd have two subclusters with two votes > each.  Neither would run, of course, since neither had quorum.E When connectivity was restored, the number of votes in each is a tie. C The tie is broken by number of nodes.  The two node subcluster wins A that so the 1 node subcluster with two votes (VELO) would reboot.   F Note that this configuration really isn't more useful than giving VELOG 1 vote and the others 0 since VELO must be up to have a running cluster E with your configuration.  I'd suggest giving each system 1 vote each, A you'd have a functional cluster with any 2 nodes up.  Unless this & configuration is temporary, of course.   ------------------------------   Date: 27 Dec 2006 07:07:14 GMT From: MI5Victim@mi5.gov.uk) Subject: MI5 Persecution: BBC h2g2 online % Message-ID: <m06112708012675@4ax.com>    BBC h2g2 16/Sept/1999   y The BBC's h2g2 website is billed as "Earth Edition of The Hitchhiker's Guide to the Galaxy!". Naturally, one of its pages 3 is about me, me, me. It starts off with the lines;    8 Every now and then something comes along which makes you8 think "Ah! That's what the Internet is for!" Mike Corley is one such case.   P The page itself can be found at http://www.bbc.co.uk/dna/h2g2/alabaster/A164404   w The user comments linked from h2g2 are less than respectful, but given the point of view my website promulgates that is x understandable. It's true that without the communications ability afforded by the internet, which came along at just thev right time, my complaints would be getting nowhere. But that's a two edged sword also; miniature cameras only appearedx around 1990, so the advantage of technology is up to the adopter. Secret service intrigues, and on the other side of the} coin human paranoia, have both been going on for centuries; what changes now is an increased technologisation of both sides.     187      --  = Posted via a free Usenet account from http://www.teranews.com    ------------------------------   Date: 27 Dec 2006 06:08:51 GMT From: MI5Victim@mi5.gov.uk3 Subject: MI5 Persecution: their methods and tactics % Message-ID: <m06112707030322@4ax.com>    -=-=-=-=-=-=-=-=-=-=-=-=-=-=-= -= MI5: methods and tactics -= -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=  D They deliberately set out to harass in a way that would resemble theH symptoms of schizophrenia, so that any report of the harassment would beH taken as indicating mental illness and "treated" accordingly. They neverI show their own faces; they only work through proxies, in the media, among J the public, and by manipulating people in the workplace. Since they do notD declare their identity there is no evidence to initiate legal actionE against the security services or anyone else. The only people you can I prosecute are the proxies and they will deny knowledge of any conspiracy.   E By repeatedly humiliating and abusing the victim, they induced mental E illness. This is the worst form of human rights violation: making any I statement of the harassment appear to be symptomatic of the illness which K they cause through the harassment. That this can happen, and people collude % by silence, is absolutely horrifying.   K From the beginning in June 1990 they set a pattern of harassment which they H have followed without change for the last six years. They  paint me as aK "threat" to which people must "react" (shades of Nazi persecution methods), F while simultaneously portraying their hate campaign on which they have< spent over a million pounds of taxpayers money as a "joke".  K The MI5 that breaks the law with the silent complicity of the police is the J same agency that is now seeking a role in the fight against crime. PerhapsJ the real joke is the proposed involvement in the implementation of justiceK of an organisation which commits criminal acts with secrecy and disinterest  for the legal process.   187      --  = Posted via a free Usenet account from http://www.teranews.com    ------------------------------  % Date: Wed, 27 Dec 2006 09:16:06 +0000 # From: Jeff Campbell <n8wxs@vcn.com> 6 Subject: Re: Need help with DHCP Client & Name servers2 Message-ID: <1167211049_4227@sp6iad.superfeed.net>   Ken.and.Ann@gmail.com wrote: > All, > H >    I've got a PWS 600au running VMS 7.3-1 and TCP/IP Services V5.3-ECO > 2, a hobbiest + > system I'm trying to put on the internet.  > I >    I have a DSL modem  from my provider (Verizon) and a LINKSYS BEFSR41 
 > (4-port)G > "EtherFast Cable/DSL Router)".   Being unable to get the DSL modem to  > talk DHCP toF > the PWS directly, I installed the LINKSYS (which had been previously > owned and therefore E > previously configured...I don't have a CD with a browser on it so I  > can't even configure theD > LINKSYS at this point...) and DHCP "got happy".  Except for DNS... > @ >    I should also note here that I've done the absolute minimum  > configuration on this "virgin"G > system: following the docs, I simply did a @Sys$Startup:Tcpip$Startup  > and was prompted@ > whether I wanted to configure DHCP Client, and I answered Yes. > H >    Very similar to the thread  "DHCP client - Help configuring.", from > May 6, 2006, in which G > Mark (shofu...@yahoo.com.au) was trying to configure a DS10 as a DHCP  > client, but under E > VMS 8.2 and TCPIP v5.5-11, I get very similar log file entries.  In  > particular, the DHCP startup/ > executes a command file that looks like this:  >  > $ SET NOON" > $ SET PROCESS/PRIV=(SYSPRV, ???)# > $ TCPIP SET NAME/SYSTEM  /DISABLE & > $ TCPIP SET NAME/SYSTEM  /NOSERVER=*9 > $ TCPIP SET NAME/SYSTEM  /SERVER=(ip-addr-1, ip-addr-2) " > $ TCPIP SET NAME/SYSTEM  /ENABLE > $ EXIT > H > (The above from my notes, I don't have access to the internet from the > system yet, duh!)  > D > The TCPIP SET NAME/SYSTEM /ENABLE fails (like Mark's did) with the > follow error:  > ; > %TCPIP-E-NAMEERROR, error processing name service request 8 > -TCPIP-E-INSBINDDATA, insufficient data to enable BIND > H > Like, what's it missing???  Similar to Mark's example, TCPIP SHOW NAME > lists the H > two ip addresses supplied by the DHCP request, and similarly, it shows > the domainC > as misconfigured (here's a cut-n-paste from Mark's post, the only  > difference is in the > pair of ip address:  >  > $ TCPIP SHOW name  >  >  > BIND Resolver Parameters >  >  >  Local domain: * Mismatch *  >  > 	 >  System  >  >   >   State:     Started, Disabled >  >  >   Transport: UDP >   Domain:  >   Retry:     4 >   Timeout:   4( >   Servers:    220.233.0.4, 220.233.0.3! >   Path:       No values defined  >  > 
 >  Process >  >  >   State:     Disabled  >  >  >   Transport: >   Domain: 
 >   Retry: >   Timeout: >   Servers:	 >   Path:  >  >  > I > I'm a little desperate at this point.  I can't ftp anywhere to download  > a browser (Mosaic or CSWB)H > because I can't resolve the names.  (I can ping the two name server IP > addresses, so I'm pretty > sure I'm getting "out".) > 1 > Any helpful advise will be greatly appreciated.  > G > Note: I can't do anything to configure the LINKSYS without a browser,  > so I need something + >          that I can do from the VMS side.  >  >     Thanks, Ken   F I too am running a PWS600au, VMS 7.3-1, TCPIP 5.3 ECO-2. I use a fixed@ addresses on network 10 for the machine but over the weekend andE Christmas I tried using the DHCP client. My DSL router is a Zoom ADSL  X6.   D The problem I ran into is that the Zoom ADSL acts as a DNS relay. ItG supplies it's own IP address as the name server address during the DHCP D configuration exchange. This address is obviously not the address of+ the BIND server I want for my local domain.   C When I configured the BIND resolver using TCPIP$CONFIG, I specified G the localhost, 127.0.0.1, as the server since I intended (as I normally B do) to run the BIND server locally. When I started TCPIP, the DHCPF client correctly picked up the leased IP address, the subnet mask, andA gateway address. It also picked up the domain and the name server  IP address.   & If you look at the log file I captured  % <http://www.n8wxs.com/web/pass_2.txt>   > you'll see (near the bottom 8-) that the DHCP client correctly@ sets the node name, as600.n8wxs.com. Also the DHCP log file does4 not complain about the BIND resolver being disabled.  = But, a $TCPIP SHOW HOST does complain about not being able to @ find the zone information (since of course the Zoom doesn't have it!)   The log file  & <http://www.n8wxs.com/web/pass_2a.txt>  A shows what happens when the machine is rebooted and TCPIP started  manually with no changes.   B The DHCP client again sets sets the various IP addresses including? the name server address. But, this time the node name is set to < unknown.n8wxs.com, and the software complains about the BIND resolver being disabled.  = For both of these experiments 8-) I removed (by renaming) the > client.pcy file in SYS$SPECIFIC:[TCPIP$DHCP]. According to the= fine manuals, the DHCP client will use default values for the B items requested in the client.pcy file if the file does not exist.  ? Obviously, the node name "unknown" is not the default contained  in the HOSTNAME. file!  ? The TCPIP permanent configuration has my original assignment of A the localhost as the bind resolver. The DHCP client replaces that ' address with the Zoom supplied address.   @ Why doesn't the DHCP client *add* the Zoom's name server address? to the name server list instead of replacing the existing local  host entry??  < Likewise, the Zoom has no idea of the Alpha's node name. The= DHCP client does not send a host name to use to the Zoom. So, 5 why didn't the DHCP client read the HOSTNAME. file in < SYS$SPECIFIC:[TCPIP$DHCP] as it did before? The Zoom did not? supply a hostname either time as can be seen in the DHCPSHOWDHC  outputs.   Puzzled,   8-)  8-)    Jeff          Q ----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==---- S http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups K ----= East and West-Coast Server Farms - Total Privacy via Encryption =----    ------------------------------    Date: 27 Dec 2006 05:17:29 -0800$ From: "AEF" <spamsink2001@yahoo.com>J Subject: Re: Nonpaged pool expansion due to burst of OPCOM DECnet messagesB Message-ID: <1167225449.590418.205280@73g2000cwn.googlegroups.com>   Bob Gezelter wrote:  > AEF wrote:I > > Recently my company's network in its NY office went mostly down. This F > > caused a major flood of OPCOM messages from DECnet about adjacencyK > > being up and down and such. The result was that nonpaged pool expansion I > > occurred in varying amounts on many of my VAX systems. On nine of the @ > > systems the expansion went all the way to the max: NPAGEVIR. > > K > > I found an old comp.os.vms thread saying that a burst of OPCOM messages * > > can rapidly exhaust nonpaged pool. OK. > > ( > > Questions: (I'm running VMS 6.2 VAX) > > F > > 1.) Is there a patch? I searched through all the release notes for/ > > non-paged and didn't see anything relevant.  > > I > > 2.) Anything to do except to reboot to reclaim locked-up memory? IOW, 9 > > is there a way to protect against this, or something?  > > H > > 3.) I thought running out of nonpaged pool was bad and could cause aF > > system crash. None of my systems crashed from this. What saved me? > >  > > Thanks.  > >  > > AEF  >  > AEF, > G > Running out of pool can indeed cause a crash, but if the shortfall is  > momentary, one can be lucky. > E > In your case, as described in the posting, it would appear that the H > burst of messages caused the pool to expand, and luckily the transientI > condition of full pool did not obstruct anything important [this time].  > I > I doubt that there is anything that can be done. On the other hand, are F > the systems affected by this pool expansion operating in an abnormalG > way? If normal operation is ongoing, without ill effects, there is no  > operational problems.  > & > - Bob Gezelter, http://www.rlgsc.com  D Everything's running fine, thanks. My NY systems are lightly loaded,* though I will do some reboots soon anyway.   Thanks   AEF    ------------------------------  + Date: Wed, 27 Dec 2006 16:32:16 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) J Subject: Re: Nonpaged pool expansion due to burst of OPCOM DECnet messages( Message-ID: <emu76g$ld1$1@pcls4.std.com>  H > > 3.) I thought running out of nonpaged pool was bad and could cause aF > > system crash. None of my systems crashed from this. What saved me? > G > Running out of pool can indeed cause a crash, but if the shortfall is  > momentary, one can be lucky.  J Whether the system crashes due to the lack of NPP, depends on exactly whatJ fails to get the memory.  In drivers, there are places where if the driverH fails to get pool, it will fork and wait and try again in a moment, withI the idea the lack of pool is transient.  However, there are other places, H even in the same driver, where it can't back out, if it can't get pool, H there is nothing it can do other than to crash.  And there may be placesE where code fails to check the status of pool allocation, as this code E continues it will probably fail with an access violation.  This would > be a bug, of course, since other bad things could also happen.   ------------------------------   End of INFO-VAX 2006.713 ************************                                                                                                                                                                                        