From: Robert Hoffmann (bobx@servermasters.com)
Date: Tue Feb 25 2003 - 13:07:29 EST
Hello guys;
I have got the following problem.
On a dual CPU x86 box, I am running Solaris 8 , and one of the CPU's
seems to be doing most of the work, and in that cpu, the "ithr" value
doing an mpstat, is always zero. (see mpstat 5 output below)
This is a Compaq Proliant DL380, with 2 cpu's, and I have Solaris 8 X86
installed, and the LATEST recommended patches (the feb 2003 one) , and
also the ATA patch for x86.
The machine uses the Compaq integrated disk array, with SCSI disks. The
ATA IDE thingy is only used for the CDROM.
I have seen that the ISY0040 (the ATA chip) is bound to CPU 1, during
boot, so I thought that was the problem.
I have 6 other machines with the EXACT same configuration, and they dont
have any problems.
Am I missing something ? Or should I throw the machine from the window ?
bash-2.03# mpstat 5
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 89 3 14 248 122 44 0 3 1 0 131 0 2 3 95
1 46 5 27 115 0 59 0 5 1 0 150 0 4 3 94
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 1 0 0 391 291 40 0 1 0 0 1 0 0 0 100
1 0 0 0 100 0 396 0 1 0 0 668 0 17 0 83
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 404 303 330 0 0 0 0 527 0 2 0 98
1 0 0 0 100 0 121 0 1 0 0 164 0 19 0 81
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 388 287 385 0 1 0 0 658 0 2 0 98
1 0 0 0 100 0 39 0 1 0 0 1 0 24 0 76
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 352 252 316 0 1 0 0 520 0 1 0 99
1 0 0 0 100 0 36 0 1 0 0 3 0 20 0 80
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 141 3 5 405 299 249 1 5 2 0 985 1 1 5 93
1 115 3 6 105 0 225 1 5 1 0 622 1 22 4 73
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 0 0 0 330 230 27 0 0 0 0 3 0 0 0 100
1 0 0 0 100 0 263 0 0 0 0 411 0 16 1 84
CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl
0 1 0 2 205 105 17 0 0 0 0 3 0 0 0 100
1 0 0 0 100 0 14 0 0 0 0 0 0 1 0 99
thanks a lot for your help if you can help.
hehe
Robert Hoffmann
ServerMasters Corp.
sunmanagers-request@sunmanagers.org wrote:
>Send sunmanagers mailing list submissions to
> sunmanagers@sunmanagers.org
>
>To subscribe or unsubscribe via the World Wide Web, visit
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
>or, via email, send a message with subject or body 'help' to
> sunmanagers-request@sunmanagers.org
>
>You can reach the person managing the list at
> sunmanagers-admin@sunmanagers.org
>
>When replying, please edit your Subject line so it is more specific
>than "Re: Contents of sunmanagers digest..."
>
>
>Today's Topics:
>
> 1. Summary : where is x86. (mike.salehi@kodak.com)
> 2. DNS root caching? (Mike's List)
> 3. How do I read a tape tarred on HPUX (Wimmer, Jeffrey)
> 4. UPDATE: DNS root caching? (Mike's List)
> 5. SUMMARY: 2 default routes and vpn (bsd unix)
> 6. SUMMARY: 2 default routes and vpn (bsd unix)
> 7. X, CDE and Tarantella over ssh (Angel Alejandro Vega Soto)
> 8. FTP error: Can't build data connection.... (Rafael Angarita)
> 9. apache (Ian Pease)
>
>--__--__--
>
>Message: 1
>Subject: Summary : where is x86.
>To: sunmanagers@sunmanagers.org
>From: mike.salehi@kodak.com
>Date: Tue, 25 Feb 2003 09:45:37 -0500
>
>Thanks folks, Too many replies to list but real thanks the answer is
>
>http://wwws.sun.com/software/solaris/binaries/get.html
>
>Mike Salehi Mike.Salehi@kodak.com (585)724-5445
>
>--__--__--
>
>Message: 2
>Date: Tue, 25 Feb 2003 09:01:05 -0600 (CST)
>From: "Mike's List" <mikelist@sky.net>
>To: sunmanagers@sunmanagers.org
>Subject: DNS root caching?
>
>~$ nslookup www.nfl.com ns1.sky.net
>Server: ns1.sky.net
>Address: 209.90.0.2
>
>Non-authoritative answer:
>Name: www.nfl.com
>Addresses: 64.253.196.251, 64.253.196.252
>-----
>~$ nslookup www.nascar.com ns1.sky.net
>Server: ns1.sky.net
>Address: 209.90.0.2
>
>*** No address (A) records available for www.nascar.com
>
>
>Does the above indicates there's a problem w/my root caching? I can do
>nslookup on most sites, but not on some sites, ie. displays on nascar.com
>What do I need to look/check for? named.local? named.root?...been at this
>for a couple of days now. Running Bind 9.2.1.
>
>Thanks.
>
>
>- Mike
>
>
>[ Packages: ftp.patriots.net (sparc/x86) | www.ondruska.com (x86) ]
>
>--
>
> "Those willing to give up a little liberty for a little security
> deserve neither security nor liberty."
> -- Benjamin Franklin
>
>--__--__--
>
>Message: 3
>From: "Wimmer, Jeffrey" <Jeffrey.Wimmer@northropgrumman.com>
>To: "'sunmanagers@sunmanagers.org'" <sunmanagers@sunmanagers.org>
>Subject: How do I read a tape tarred on HPUX
>Date: Tue, 25 Feb 2003 07:25:43 -0800
>
>Hello fellow gurus,
>
>I have a tape I believe was tarred on an HPUX 9.05 system. I would like to
>read this tape on my Solaris 8 system. Performing a tar tv gives me the
>following:
>
># tar -tv
>tar: directory checksum error
>
>performing a tar xv gives me the same result. Not sure if I can use dd to
>extract the data, but when I do the following, I get the error given too.
>
># dd if=/dev/rmt/0m conv=swab | tar -xvf -
>read: Not enough space
>0+0 records in
>0+0 records out
>tar: blocksize = 0
>
>Help please!!!
>
>Thanks in advance
>Jeff Wimmer.
>
>--__--__--
>
>Message: 4
>Date: Tue, 25 Feb 2003 09:35:13 -0600 (CST)
>From: "Mike's List" <mikelist@sky.net>
>To: sunmanagers@sunmanagers.org
>Subject: UPDATE: DNS root caching?
>
>Here's another strange thing, I can now do all nslookup at my nameserver
>machine --then my clients can do nslookup on that particular domain just
>fine. But if I don't do "nslookup nascar.com" then my clients can't do
>nslookup nascar.com. So it seems the NS doesn't go to the root server
>somehow?
>
>I've done dig @a.root-servers.net . ns and pipe it into named.root file,
>I've stop/start named. Bind is running w/permission of named:named not
>as root.
>
>
>- Mike
>
>
>---------- original message ----------
>
>~$ nslookup www.nfl.com ns1.sky.net
>Server: ns1.sky.net
>Address: 209.90.0.2
>
>Non-authoritative answer:
>Name: www.nfl.com
>Addresses: 64.253.196.251, 64.253.196.252
>-----
>~$ nslookup www.nascar.com ns1.sky.net
>Server: ns1.sky.net
>Address: 209.90.0.2
>
>*** No address (A) records available for www.nascar.com
>
>
>Does the above indicates there's a problem w/my root caching? I can do
>nslookup on most sites, but not on some sites, ie. displays on nascar.com
>What do I need to look/check for? named.local? named.root?...been at this
>for a couple of days now. Running Bind 9.2.1.
>
>Thanks.
>
>
>- Mike
>
>
>[ Packages: ftp.patriots.net (sparc/x86) | www.ondruska.com (x86) ]
>
>--
>
> "Those willing to give up a little liberty for a little security
> deserve neither security nor liberty."
> -- Benjamin Franklin
>_______________________________________________
>sunmanagers mailing list
>sunmanagers@sunmanagers.org
>http://www.sunmanagers.org/mailman/listinfo/sunmanagers
>
>--__--__--
>
>Message: 5
>From: "bsd unix" <bsdunix@mail.com>
>To: sunmanagers@sunmanagers.org
>Date: Tue, 25 Feb 2003 10:47:52 -0500
>Subject: SUMMARY: 2 default routes and vpn
>
>I got no replies with other ideas. I tried Casper's ipfilter trick
>and it works -- packets route through the correct cards now.
>
>Alas, the combination of the IPsec tunnel and ipfilter are now causing
>kernel stack overflows, even with lwp_default_stksize set to 0x8000.
>
>Additionally, Sunsolve has conflicting information on what it should
>be set to: infodoc 20151 says "increasing stack sizes above 16k is a
>wholesale waste of memory" yet infodoc 40722 recommends increasing
>"the value to the next higher setting...If it is currently 16k, make
>it 24k."
>
>Based on the description of rpcmod:svc_run_stksize it would seem to have
>no impact on this setup, but I will try adding that anyway.
>
>Thanks to Casper Dik for having posted his ipfilter trick.
>
>
>----- Original Message -----
>From: "bsd unix" <bsdunix@mail.com>
>Date: Sat, 22 Feb 2003 19:41:49 -0500
>To: sunmanagers@sunmanagers.org
>Subject: 2 default routes and vpn
>
>
>
>>I have a system which has 2 default routes and is one end of a vpn.
>>Snoop shows that sometimes the system sends packets with the source
>>address of one card out the other. There is one network per card and
>>one default route per card.
>>
>>What I am reading suggests that Solaris is using round-robin when
>>sending the packets out. However, this causes some vpn packets to go
>>out the wrong interface, and thus the vpn has intermittent outages.
>>I saw a SUMMARY where Casper gave an example of using ipfilter to
>>force routing of the right network traffic out of the right card, but
>>was hoping there would be some kernel parameter or tweak to do this
>>natively.
>>
>>Apparently ip_strict_dst_multihoming is of no help because that only
>>works on inbound connections.
>>
>>Thanks in advance for any help or pointers.
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:25:53 EDT