From chan.playpool at gmail.com  Fri Aug  1 10:30:42 2008
From: chan.playpool at gmail.com (the bx)
Date: Fri, 1 Aug 2008 10:30:42 -0400
Subject: SUMMARY Re: Help with jumpstarting V440
Message-ID: <ada67b5d0808010730k3950eb2bl3fd9c6b052e31686@mail.gmail.com>

really a non-summary.

Gurus:

I'm still experiencing the problem.  I had other issues with the JS
server too.  One of which is related to moving the server to a
different IP.  I neglected to change
some settings after the move.  I thought of OS version, arch, etc too
as some had suggested.  This server/boot server/OS rev combo has been
responsible for
the births of many similar Netra V440's in the past.  I'll stick with
troubleshooting and give updates.

Thanks to the following Gurus for responding,

John England
Vikas Sharma
Matthew StierLoading...	
JayJay Florendo

Chan


On Thu, Jul 31, 2008 at 12:58 PM, the bx <chan.playpool at gmail.com> wrote:
> Gurus:
>
> I got a few hits off google but I'm still lost.
>
> I tried to jumpstart Sol10 from a Sol 8 boot server and got this,
>
> Rebooting with command: boot net - install
> Boot device: /pci at 1c,600000/network at 2  File and args: - install
> /pci at 1c,600000/network at 2: 100 Mbps half duplex link up
>
> Requesting Internet address for 0:14:4f:d6:87:d8
> krtld: load_exec: fail to expand cpu/$CPU
> /platform/SUNW,Netra-440/kernel/misc/sparcv9/platmod symbol
> plat_lgrp_latency multiply defined
> /platform/SUNW,Netra-440/kernel/misc/sparcv9/platmod symbol
> plat_lgrp_cpu_to_hand multiply defined
> /platform/SUNW,Netra-440/kernel/misc/sparcv9/platmod symbol
> plat_lgrp_init multiply defined
> /platform/SUNW,Netra-440/kernel/misc/sparcv9/platmod symbol
> plat_build_mem_nodes multiply defined
> krtld: error during initial load/link phase
> panic - boot: exitto64 returned from client program
> Program terminated
>
> The install server is a sol10, x86.  I'm trying to install 05/2008
> version of 10.
>
> Please dispense wisdom and thanks!
>
> Chan
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From sunhux at gmail.com  Mon Aug  4 05:32:50 2008
From: sunhux at gmail.com (sunhux G)
Date: Mon, 4 Aug 2008 17:32:50 +0800
Subject: Summary: V880 slow with scan rates in thousands
Message-ID: <60f08e700808040232p73ddfd04h4930f90ef138a1ef@mail.gmail.com>

Thanks to Richard Skelton for the instantaneous response :
  Top in Solaris can sort by size:-
  Press o type size
  also better to use prstat -s size

& the SIZE column in prstat/top showed that among the largest
values of SIZE are owned by oracle processes.

Initially stopping all applications did not help bring down the
scan rate;  after stopping all the 6 Oracle instances, the scan
rate came to 0.

We're not able to reproduce this problem after all the 6 Oracle
instances & the application are started up again.


Thanks to Maciej for responding within 15 minutes.  We'll need
to do a post-mortem to see which process chews up the memory
- difficult to justify to add RAM unless the problem is persistent
& we can pinpoint the culprit.


Thanks
U





On 8/4/08, sunhux G <sunhux at gmail.com> wrote:
>
> Hi,
>
>
>
> Just today, the scan rates of our V880 jumped to thousands
> (used to be 0).  CPU utilization is 63-87% idle :
>
> vmstat 2:
>  0 1 0 18565856 59753 395 172 582 30 54 40184 171 20 0 0 0 1231 1202 892 4
> 7 89
>  0 6 0 14795056 12778  93 215 1003 311 1582 32552 2811 24 0 0 0 748 556
> 1148 2 5 93
>  0 3 0 14795056 12777  86 205 970 448 1608 26376 1975 7 0 0 0 791 936 1154
> 3 5 91
>  0 3 0 14793024 12694 152 546 3073 975 3300 113240 4846 17 0 0 0 920 1223
> 1228 5 12 83
>  0 3 0 14795648 13065 114 265 1166 100 100 91728 0 21 0 0 0 651 634 1193 2
> 3 95
>  0 3 0 14795648 12783  58 178 907 585 1091 74304 2316 7 0 0 0 709 600 1055
> 2 5 93
>  7 3 0 14794864 12804 148 350 2838 957 2199 115176 30854 18 0 0 0 751 658
> 1177 3 18 79
>
>  0 4 0 15158376 128016 153 302 1009 1555 1627 44648 16005 4 0 0 0 591 575
> 1075 3 10 87
>  0 3 0 15158376 128584 328 479 990 3485 3616 36168 30039 20 0 0 0 660 468
> 1127 2 13 85
>  0 3 0 15158376 128344 384 494 784 3813 3916 29304 26749 14 0 0 0 615 517
> 1030 3 13 84
>  0 3 0 15158376 127144 350 491 884 3590 3821 23744 22907 7 0 0 0 600 479
> 1027 2 12 86
>
> iostat 2:
>
>    tty        md3          md104         md105         md106           cpu
>  tin tout kps tps serv  kps tps serv  kps tps serv  kps tps serv   us sy wt
> id
>    0    1 480  20    8    0   0    7    0   0   13    2   0   17    4  7  0
> 89
>    0  117  85   7    9    0   0    0    0   0    0    0   0    0    5  3  0
> 92
>    0   40 121  15   10    0   0    0    0   0    0    0   0    0    2  3  0
> 95
>
> top output:
>
> last pid: 26802;  load averages:  0.28,  0.38,
> 0.54                                    17:01:49
> 165 processes: 163 sleeping, 1 zombie, 1 on cpu
> CPU states:     % idle,     % user,     % kernel,     % iowait,     % swap
> Memory: 8192M real, 5703M swap in use, 17G swap free
>
>   PID USERNAME THR PRI NICE  SIZE   RES STATE   TIME    CPU COMMAND
> 26802 root       1   0    0 2392K 1864K cpu0    0:00  1.23% top
> 18999 root       1  59    0 1240K  456K sleep  10:18  0.28% dd
> 19635 root       1  59    0 1240K  456K sleep  10:11  0.28% dd
> 26800 oracle10   1  59    0  524M  507M sleep   0:00  0.23% oracle
>  7574 oracle10   1  59    0  579M  556M sleep   0:37  0.11% oracle
>  7860 oracle10   1  59    0 1329M 1304M sleep   0:42  0.11% oracle
>  7570 oracle10   1  59    0  580M  556M sleep   0:42  0.10% oracle
>  7720 oracle10   1  59    0  524M  500M sleep   0:42  0.10% oracle
>  7504 oracle10   1  59    0 1031M 1008M sleep   0:07  0.10% oracle
>  7776 oracle10   1  59    0 1329M 1304M sleep   0:43  0.09% oracle
>
>
>
> I know in Linux, the top command can sort by CPU/memory/swap
>
> usage (just by pressing F  & select the sort column/key) but the
>
> top utility do not have this in Solaris
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From earlysame55 at gmail.com  Tue Aug  5 16:10:15 2008
From: earlysame55 at gmail.com (Unix Administrator)
Date: Wed, 6 Aug 2008 00:10:15 +0400
Subject: SUMMARY:Kernel ID
Message-ID: <2a81355a0808051310n5bab3c49r9ec19771b6258744@mail.gmail.com>

Dear all,

Firstly, Rajiv Gunja thanks for your reply to look in the
patchdiag.xref in the EIS CD. I managed to find the EIS CD version
from the /var/sun directory, there is a EIS CD log. It had the info i
needed.

Thanks
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From admitriev at mentora.biz  Tue Aug  5 17:36:17 2008
From: admitriev at mentora.biz (Andrey Dmitriev)
Date: Tue, 5 Aug 2008 17:36:17 -0400
Subject: SUMMARY (sort of): horrible ZFS performance on a pool of 2
	LUNs, awesome with 1 LUN
Message-ID: <973755.15511217972177269.JavaMail.root@mail.mentora.biz>

Well, after eliminating every possible hardware issue, the issue turned out to be with software, particularly with utilization of the underlying ZFS file system. As soon as we deleted about 2TB worth of data, the file system started performing fine again.
 
we deleted beast/customer1/db file system (had it on tape, just like others), and right away were back to 150-300megs per sec.
 
If anyone has a clue as to why this might've been an issue, or advise some ZFS related mailing list, I'd greately appreciate it.
 
beast                 130G   37K  130G   1% /mnt/backup1
beast/customer1         130G   29K  130G   1% /mnt/backup1/customer1
beast/customer1/bacula  222G   93G  130G  42% /mnt/backup1/customer1/bacula
beast/customer1/db      2.0T  1.8T  130G  94% /mnt/backup1/customer1/db
beast/customer1/fs      2.1T  1.9T  130G  94% /mnt/backup1/customer1/filesystem
beast/customer5      130G   29K  130G   1% /mnt/backup1/customer5
beast/customer5/bacula
                      221G   92G  130G  42% /mnt/backup1/customer5/bacula
beast/customer5/db   130G   25K  130G   1% /mnt/backup1/customer5/db
beast/customer5/fs   172G   42G  130G  25% /mnt/backup1/customer5/filesystem
beast/bacula          130G   15M  130G   1% /mnt/backup1/bacula
beast/bacula/spool    130G   34K  130G   1% /mnt/backup1/bacula/spool
beast/customer6          130G   29K  130G   1% /mnt/backup1/customer6
beast/customer6/bacula   210G   81G  130G  39% /mnt/backup1/customer6/bacula
beast/customer6/db       3.7T  3.6T  130G  97% /mnt/backup1/customer6/db
beast/customer6/fs       130G   25K  130G   1% /mnt/backup1/customer6/filesystem
beast/customer2         133G  3.6G  130G   3% /mnt/backup1/customer2
beast/customer2/bacula  1.5T  1.4T  130G  92% /mnt/backup1/customer2/bacula
beast/customer2/db      194G   65G  130G  34% /mnt/backup1/customer2/db
beast/customer2/fs      221G   92G  130G  42% /mnt/backup1/customer2/filesystem
beast/customer4         130G   29K  130G   1% /mnt/backup1/customer4
beast/customer4/bacula  1.3T  1.2T  130G  90% /mnt/backup1/customer4/bacula
beast/customer4/db      1.6T  1.5T  130G  92% /mnt/backup1/customer4/db
beast/customer4/fs      130G   25K  130G   1% /mnt/backup1/customer4/filesystem
beast/customer3    130G   26K  130G   1% /mnt/backup1/customer3
beast/customer3/bacula
                      2.8T  2.6T  130G  96% /mnt/backup1/customer3/bacula

Original Post:
                                           capacity     operations    bandwidth
pool                                     used  avail   read  write   read  write
--------------------------------------  -----  -----  -----  -----  -----  -----
beast                                   14.1T   366G      0    155      0  3.91M
  c7t6000402002FC424F6CF5317A00000000d0  7.07T   183G      0     31      0  16.2K
  c7t6000402002FC424F6CF5318F00000000d0  7.07T   183G      0    124      0  3.90M 

I get pretty consistent results like this.. i can only write to the pool at about 3megs right now.. i used to at about 300-400MB/sec

each member is a 9 1TB disk RAID5, members are not mirrored

I have another group that I created on the same array (NexSAN SATABeast). using only 2 disks (mirror). I am able to push them to 60megs, which is fine.

We have tried rebooting switches, setting all ports to 2Gb hard, eliminating controllers (each controller presents all LUNs), eliminating ports on the fibre card, direct attaching the machine to the array, yet I am consistently getting the same (crappy) results on one LUN, and decent on another.

Interestingly I see reads in the 30-40megs per member all the time, but writes suck consistently. 

We had a network maintenance the day before, which was also rolled back.

Anyone has _any_ clue on how to troubleshoot this further?


 
FOLLOW UP:
 
I do see a problem (this was done with 30secs) iostat intervals

                 extended device statistics
device    r/s    w/s   kr/s   kw/s wait actv  svc_t  %w  %b
    0.0    6.3    0.0    3.3  0.0  0.0    0.0    0.6   0   0 c7t6000402002FC424F6CF5318F00000000d0s0
    0.0   25.1    0.0  805.3  0.0  0.0    0.0    1.4   0   2 c7t6000402002FC424F6CF5317A00000000d0s0


    1.4    6.9    3.4    3.3  0.0  0.0    0.0    0.5   0   0 c7t6000402002FC424F6CF5318F00000000d0s0
    1.4   26.1    3.4  822.5  0.0  0.0    0.0    1.4   0   2 c7t6000402002FC424F6CF5317A00000000d0s0

   73.0   11.6 4476.2  165.0  0.0  2.9    0.0   34.1   0  17 c7t6000402002FC424F6CF5318F00000000d0s0
   76.4   19.3 4727.3  487.2  0.0  2.6    0.0   27.1   0  17 c7t6000402002FC424F6CF5317A00000000d0s0

However, I do not understand why ZFS is making writes in such a lopsided manner. 
E.g. why are number of writes equal to the LUNs, yet the # of KB is substantially different

Also, we did try swapping cables.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From stepchung at gmail.com  Wed Aug  6 12:02:33 2008
From: stepchung at gmail.com (Stephanie C)
Date: Wed, 6 Aug 2008 09:02:33 -0700
Subject: SUMMARY: PLEASE READ - Oracle DB on ZFS Version
Message-ID: <6bdfa8ce0808060902y75adb2bcy66b543621843f101@mail.gmail.com>

I would like to withdraw my question due to one gentlemen (hike <
mh1272 at gmail.com>) who is giving me a hard time about my question. I treat
every questions on this list that subjects to SUN related with respect, even
simple or stupid questions. I don't give people hard time because their
questions, I ignore them. Since this gentlemen gave me a hard time about
this, please ignore my question. Thank you very much for all replied my
questions, 18 educational institutions are running Oracle on ZFS with no
problem. Thanks again.

Note: I will have to be careful about the questions on this list so people
do not attack me.

Stephanie


Original question:

Hello- I posted a question earlier about this issue. I got very good
responses and already summarized the result on the list (summary again
below). That is not enough to convince my boss to go with ZFS for Oracle
database. We are an higher Education intuition. He wants to know if any
colleges or universities out there run Oracle DB on ZFS and any problem or
issue with it. He does not want 'we are the first education institution to
do this'. Please let me know if you are an education institution running
oracle db on zfs. Thank you.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From jclark42 at gmail.com  Thu Aug  7 10:39:54 2008
From: jclark42 at gmail.com (Joshua Clark)
Date: Thu, 7 Aug 2008 10:39:54 -0400
Subject: Summary: Solaris 10 Cluster- nfsd Generating Huge CPU Load
Message-ID: <6f4407960808070739g7878cb0ao765c7555c8b0eac3@mail.gmail.com>

This problem was caused by a few misbehaving HP-UX NFS clients. I
found Solaris bug 6538387, which describes an incompatibility between
HP-UX NFS v3 clients and Solaris NFS servers running ZFS.  The
solution was to umount the NFS file system from the five Hp-UX clients
and re-mount it using NFSv2. The CPU load on the NFS server immediatly
dropped to near zero and performance is now fine on all NFS clients.

Thanks everyone for your help nailing this problem down.

Josh

On Thu, Aug 7, 2008 at 7:56 AM, Joshua Clark <jclark42 at gmail.com> wrote:
> Thank you all for your responses so far. I have made significant
> progress with this issue. I believe I have narrowed the problem to the
> few HPUX NFS clients on our network. We have 120+ NFS clients, a mix
> of Solaris, Linux, BSD, HP-UX, AIX, and some others like VMS, etc.
>
> Looking at "nfsstat -s" I see a huge number of SETATTR calls:
>
> bash-3.00# nfsstat
> <snip>
> Server nfs:
> calls            badcalls
> 1488894256  3222
> <snip>
> Version 3: (1488672395 calls)
> null           getattr                setattr                 lookup
>           access
> 2 0%        160776192 10%  1313974792 88% 10634878 0%    1092862 0%
> readlink          read                write               create
>     mkdir
> 5336 0%        871124 0%      283688 0%      60968 0%       7594 0%
> ...
>
> I ran snoop for 10 seconds on the GigE interface. It captured 30,059
> packets. Of these, 12,284 were SETATTR calls, and a total of 29,969
> were between the NFS server and our five HP-UX clients. I checked
> these systems- there is little or no activity that would warrant this
> much traffic. I believe there is some sort of problem between the
> HP-UX NFS clients and the Sun NFS server that is causing a packet
> storm which is driving the CPU usage on the NFS server.
>
> Has anyone had any problems similar to this? Any ideas where to turn next?
>
> Thanks again,
>   Josh
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From ahoesch at smartsoft.de  Mon Aug 11 07:08:36 2008
From: ahoesch at smartsoft.de (=?ISO-8859-1?Q?Andreas_H=F6schler?=)
Date: Mon, 11 Aug 2008 13:08:36 +0200
Subject: Summary: Second NIC on X2200
In-Reply-To: <DAA7B7C2-678E-11DD-9BD0-000393CA0072@smartsoft.de>
Message-ID: <D830DBDA-6795-11DD-9BD0-000393CA0072@smartsoft.de>

Dear managers,

> I have Solaris 05/08 installed on a X2200 and am trying to setup a
> second NIC without success. I have done the usual
>
> 	pico /etc/hosts
> 	pico /etc/netmasks
> 	pico /etc/hostname.nge1
>
> stuff and then did "svcadm restart network/physical". Nothing! "netstat
> -rn" still shows only one interface:
>
> Routing Table: IPv4
>    Destination           Gateway           Flags  Ref     Use
> Interface
> -------------------- -------------------- ----- ----- ----------
> ---------
> 192.168.1.0          192.168.1.100        U         1          1 nge0
> 224.0.0.0            192.168.1.100        U         1          0 nge0
> 127.0.0.1            127.0.0.1            UH        9         60 lo0
>
> I then tried to plumb the guy maually
>
> 	ifconfig nge1 unplumb
> 	ifconfig nge1 plumb 217.6.81.X netmask 255.255.255.Y up
>
> but still no avail. I get no error message but also no second interface
> in "netstat -rn".
>
> What am I missing? Is there a special trick on thi snew OS release. I
> realized that /etc/inet/ipnodes now is a link to /etc/hosts. Great! But
> this also tells me that they have modified networking stuff and my
> normal receipt no longer works! :-( Hints are greatly appreciated!

Replying to my own stupid posting. I finally tried sys-unconfig and 
while doing so saw a suspiscious message on the screen telling me that 
the ipaddress I wanted to give my second NIC is already being used by 
another machine on the network. :-( I tried another ipaddress and it 
works like a charm!

Thanks a lot!

Andreas
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From KSmith at espeed.co.uk  Tue Aug 12 11:24:49 2008
From: KSmith at espeed.co.uk (Smith, Kev)
Date: Tue, 12 Aug 2008 16:24:49 +0100
Subject: V1280 Console line wrap issue - SUMMARY
In-Reply-To: <20136350C0106746A13E689A887517A007210081@ESSHEXCL1VS1.eu.ad.espeed.com>
References: <20136350C0106746A13E689A887517A00721007A@ESSHEXCL1VS1.eu.ad.espeed.com>
	<20136350C0106746A13E689A887517A007210081@ESSHEXCL1VS1.eu.ad.espeed.com>
Message-ID: <20136350C0106746A13E689A887517A007210084@ESSHEXCL1VS1.eu.ad.espeed.com>

/sbin/autopush had 644 permissions, should have been 555 - We are
looking at adhoc OS disk corruption at the moment as we are seeing other
issues on the host as well. autopush deals with stream IO including the
console. The file was corrected and the stepping issue was fixed

FYI SUN DOC  202387 details [which has been talked about before on this
list - see "console stepping"]:
>>
Check the serial port entries in /etc/path_to_inst:

# grep serial /etc/path_to_inst

The entries should look like:

    "/pci at 1f,0/isa at 7/serial at 0,2e8" 1 "su"
    "/pci at 1f,0/isa at 7/serial at 0,3f8" 0 "su"
If these entries are incorrect or missing, then take one of actions
below:

1) If this system was just installed with JumpStart[TM], then most
likely the jumpstart
   image is corrupt. Recreate the image and jumpstart the client again.


2) If the system had been functioning normally before this problem
occurred, then
   recreate the path_to_inst file:
   # mv /etc/path_to_inst ..
   # init 0
   ok boot -a

   [Press Return to select default values for everything except
/etc/path_to_inst.
   Enter y to rebuild /etc/path_to_inst when prompted]:

   The /etc/path_to_inst on your system does not exist or is empty.
   Do you want to rebuild this file [n]  y
<<

Thanks to Ric Anderson, Juraj Lutter, JayJay Florendo & BV for their
input to the problem
http://www.espeed.com
CONFIDENTIAL: This e-mail, including its contents and attachments, if any, are
confidential. If you are not the named recipient please notify the sender and
immediately delete it. You may not disseminate, distribute, or forward this
e-mail message or disclose its contents to anybody else. Copyright and any
other intellectual property rights in its contents are the sole property of
eSpeed, Inc and its affiliates.
This e-mail was issued by eSpeed International Limited (eSpeed). eSpeed is a
limited liability company incorporated under the laws of England (company
number 3809189 and VAT registration number 577 406809). eSpeeds registered
office is at 40 Bank Street, Canary Wharf, London E14 5DW. For any issues
arising from this email please reply to the sender.
E-mail transmission cannot be guaranteed to be secure or error-free. The
sender therefore does not accept liability for any errors or omissions in the
contents of this message which arise as a result of e-mail transmission. If
verification is required please request a hard-copy version.
Although we routinely screen for viruses, addressees should check this e-mail
and any attachments for viruses. We make no representation or warranty as to
the absence of viruses in this e-mail or any attachments. Please note that to
ensure regulatory compliance and for the protection of our customers and
business, we may monitor and read e-mails sent to and from our server(s).
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Rob.McMahon at warwick.ac.uk  Wed Aug 13 05:52:01 2008
From: Rob.McMahon at warwick.ac.uk (Rob McMahon)
Date: Wed, 13 Aug 2008 10:52:01 +0100
Subject: SUMMARY: V890 and DR - swap system board
In-Reply-To: <48A06EAF.40102@warwick.ac.uk>
References: <48A06EAF.40102@warwick.ac.uk>
Message-ID: <48A2AEC1.1070607@warwick.ac.uk>

Rob McMahon wrote:
> The implication from Sun is that they can replace a system board with 
> the system live, but cfgadm doesn't show any cpu / memory modules, so 
> I'm getting suspicious.  Can we really take half the memory and CPUs out 
> of service while they replace the board, or is this going to mean 
> downtime.  How do we do this ?  (This is a pretty important server, and 
> downtime would need to be carefully scheduled.)
>   
Thanks for all your input.  I think a reasonable summary is that 
although the documentation says it's not supported, if you're up to date 
with the firmware you might get away with it, but no-one in their right 
mind would risk it.  In theory you can bodge the buttons inside to 
offline the board and /usr/platform/SUNW,Sun-Fire-V890/lib/sf880drd 
should spot the fact and migrate everything off there to allow you to 
pull the board.  If you've made up with your favorite deity.

Anyway, scheduled downtime it is.

Thanks to Scott Lawson, Dean Ross-Smith, JayJay Florendo, Tim Bradshaw, 
and Joe Fletcher.  Your input is really appreciated.

Cheers,

Rob

-- 
E-Mail:	Rob.McMahon at warwick.ac.uk		PHONE:  +44 24 7652 3037
Rob McMahon, IT Services, Warwick University, Coventry, CV4 7AL, England
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From breynolds at Cymtec.com  Wed Aug 13 15:16:58 2008
From: breynolds at Cymtec.com (Beverley Reynolds)
Date: Wed, 13 Aug 2008 14:16:58 -0500
Subject: SUMMARY - mirror disk1 will not boot
In-Reply-To: <mailman.5579.1218559003.292.sunmanagers@sunmanagers.org>
References: <mailman.5579.1218559003.292.sunmanagers@sunmanagers.org>
Message-ID: <800519794AFB8B47A201D7B8B7B1502E367BD7@cymex.Cymtec.net>

Thanks for all the replies.

Everyone said to run

installboot /usr/platform/`uname -i`/lib/fs/ufs/bootblk
/dev/rdsk/c0t1d0s0

Which I did.
But I also had to apply about 80 patches per SUN support and undo the
mirrors and rebuild everything.

I can now boot from the mirror disk.


On Fri, Aug 1, 2008 at 1:49 PM, Beverley Reynolds <breynolds at cymtec.com>
wrote:
> I set up this new T2000, sliced both disks, installed and patched the
> OS and created a mirror on the second disk.
>
> When I try to boot from disk1 I get the following error..........
>
>
>
> ===================================
> Rebooting with command: boot disk1
> Boot device: /pci at 780/pci at 0/pci at 9/scsi at 0/disk at 1 File and args:
> The file just loaded does not appear to be executable.
> {0} ok
>
>
>
> ===================================
>
> Here are the disk layouts...........
>
>
>
> c0t0d0s0
>
> Total disk cylinders available: 14087 + 2 (reserved cylinders)
>
>
>
> Part      Tag    Flag     Cylinders         Size            Blocks
>
>  0       root    wm       0 -  3823       18.56GB    (3824/0/0)
> 38913024
>
>  1       swap    wu    3824 -  5836        9.77GB    (2013/0/0)
> 20484288
>
>  2     backup    wm       0 - 14086       68.35GB    (14087/0/0)
> 143349312
>
>  3       home    wm    5837 - 10257       21.45GB    (4421/0/0)
> 44988096
>
>  4 unassigned    wm       0                0         (0/0/0)
> 0
>
>  5        var    wm   10258 - 14081       18.56GB    (3824/0/0)
> 38913024
>
>  6 unassigned    wm       0                0         (0/0/0)
> 0
>
>  7 unassigned    wm   14082 - 14086       24.84MB    (5/0/0)
> 50880
>
>
>
>
>
> c0t1d0s0
>
> Part      Tag    Flag     Cylinders         Size            Blocks
>
>  0       root    wm       0 -  3823       18.56GB    (3824/0/0)
> 38913024
>
>  1       swap    wu    3824 -  5836        9.77GB    (2013/0/0)
> 20484288
>
>  2     backup    wu       0 - 14086       68.35GB    (14087/0/0)
> 143349312
>
>  3       home    wm    5837 - 10257       21.45GB    (4421/0/0)
> 44988096
>
>  4 unassigned    wm       0                0         (0/0/0)
> 0
>
>  5        var    wm   10258 - 14081       18.56GB    (3824/0/0)
> 38913024
>
>  6 unassigned    wm       0                0         (0/0/0)
> 0
>
>  7 unassigned    wm   14082 - 14086       24.84MB    (5/0/0)
> 50880
>
>
>
> ================================
>
> This is how I created the mirror
>
>
>
>
>
> metadb -a -f -c 3 c0t0d0s7
>
>
>
> metainit -f d11 1 1   c0t0d0s0
>
> metainit    d12 1 1   c0t1d0s0
>
> metainit    d10 -m d11
>
> metaroot    d10
>
>
>
>
>
> metainit -f d21 1 1   c0t0d0s1 {this is where swap is}
>
> metainit    d22 1 1   c0t1d0s1
>
> metainit    d20 -m d21
>
>
>
> metainit -f d31 1 1   c0t0d0s5 {/var}
>
> metainit    d32 1 1   c0t1d0s5
>
> metainit    d30 -m d31
>
>
>
> metainit -f d41 1 1   c0t0d0s3 {/export/home}
>
> metainit    d42 1 1   c0t1d0s3
>
> metainit    d40 -m d41
>
>
>
> ==============================
>
> I made changes to /etc/vfstab and rebooted the machine
>
> Everything seemed to run fine.
>
> I wanted to test to ensure the mirror disk would boot when I got the
> above error.
>
>
>
> I opened a case with sunsolve but have not heard back from them.
>
>
>
> Thanks
>
#############################################################################
##
This message from Cymtec Systems, Inc. contains confidential information and
is solely for the use of the recipient(s) named above.  If you are not the
intended recipient or an agent responsible for delivering it to the intended
recipient, you are hereby notified that you have received this message in
error and that any review, disclosure, copying, distribution or use of the
contents of this message is strictly prohibited.  If you have received this
message in error, please destroy it immediately and notify Cymtec Systems,
Inc. by telephone at +1.314.993.8700 or by return e-mail.
#############################################################################
##
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From sunhux at gmail.com  Sat Aug 16 08:07:57 2008
From: sunhux at gmail.com (sunhux G)
Date: Sat, 16 Aug 2008 20:07:57 +0800
Subject: Partial Summary: ufsdump from a remote server via ssh
Message-ID: <60f08e700808160507j3d9c1735k13f808a0decebaf0@mail.gmail.com>

Thanks, not a path issue but Openssh is not there in Solaris 8.

The command is :
ufsdump 0uf - test.txt | ssh slagpo at 10.51.101.5 "cd /tmp; cat > test.dmp;
gzip test.dmp"
(tested between 2 Solaris 10 servers)

(One reply corrected my original command :
  # nohup ufsdump 0uf - /  | ssh myuserid at 10.51.101.5 "cd /ora/destdir;
   zip ufsdump.zip -"


I'll need to install from www.sunfreeware.com  Openssh (& the
pre-requisites Openssl & lastly libgcc) but after restarting (by issuing
"/usr/sbin/sshd2 restart" ), ssh now has a message :

# /usr/local/bin/ssh myuserid at 10.51.101.5 <myuserid at 10.51.101.5>
PRNG is not seeded

(Some of those suggestions from google did not resolve PRNG error
  above)

NFS share is not permitted for security reason.


Thanks again,
U

On 8/15/08, sunhux G <sunhux at gmail.com> wrote:
>
> Hi
>
> I'm exploring a solution to do "ufsdump" from a Solaris 8 server (which
> currently also has an 'ssh command not found problem') to a remote
> Solaris 10 server (which has a lot of disk space) :
>
> On Solaris 8 server :
> # ssh
> ksh: ssh:  not found
>
>
> The command below on Solaris 8 did not work (partly the
> ssh problem above & possibly I'm missing something) :
>
> # nohup ufsdump 0uf - /  | ssh myuserid at 10.51.101.5 "cd /ora/destdir; zip
> -" &
>
> where /ora/destdir  is permitted for writing by the remote user myuserid.
>
>
> What's the missing bits?
>
>
> Also, suppose instead of using the remote Solaris 10 server, can I
> substitute
> it with a Windows ssh server (which also has a lot of disk space)?
>  Problem is Windows XP supports max file size of 4Gb only if I'm not wrong
>
>
> TIA
> U
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Matthew.Stier at us.fujitsu.com  Sun Aug 17 23:53:10 2008
From: Matthew.Stier at us.fujitsu.com (Matthew Stier)
Date: Sun, 17 Aug 2008 23:53:10 -0400
Subject: SUMMARY: Has anyone seen LP options changes under Solaris 10 (05/08)?
In-Reply-To: <48A55029.7070101@uni-paderborn.de>
References: <487BD028.5000705@us.fujitsu.com>
	<48A55029.7070101@uni-paderborn.de>
Message-ID: <48A8F226.8060308@us.fujitsu.com>

Found it, and the only fix is to modify the 'interfaces' scripts to 
accept both pre and post PAPI boolean options.

With Solaris 10 05/08, Sun implemented the PAPI interface, to support 
multiple printing back ends.  (BSD/IPP/CUPS/...)

The lp command calls the papiAttributeListFromString() function to 
process each option. 

    
http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/print/bsd-sysv-commands/lp.c
    http://src.opensolaris.org/source/s?defs=papiAttributeListFromString

The papiAttributeListFromString calls a _parse_attribute_list() 
function, who job it is to process the numerous option argument 
formats.  Near the end of the function, it process boolean (key only) 
options. The first thing the boolean key code does is convert the 
boolean key into a key=value pair, with the key being set to a value of 
"true".  It then test the first two characters of the key to see if it 
begins with "no". If it does, it removes the "no" prefix, and sets the 
new key to "false".  (ie: "banner" becomes "banner=true". "nobanner" 
becomes "banner=false".)

    http://src.opensolaris.org/source/s?refs=_parse_attribute_list
    (lines 853-861)

Thus the only fix is to go through each interfaces script, and modify 
the code that process the options, to accept both 'key' and 'key=value' 
options.


Jan Dreyer wrote:
> Hi,
>
> though your message is quiet old, I would like to point out, that we 
> encountered similar problems with clients on HP-UX. Our solution was 
> to fix the scripts which send the data and make little tweaks to our 
> own scripts in .../interfaces.
>
> Did you find another solution or a hint what SUN changed?
>
> Greetings
> Jan Dreyer
>
>
> Matthew Stier schrieb:
>> I've just deployed Solaris 10 (05/08) to a number of users, and an 
>> encountering a problem with some /etc/lp/interfaces scripts I wrote 
>> years ago.  This problems has not been encountered on any previous 
>> release of Solaris 10, or 9, 8, 7, ...
>>
>> It appears that the latest release of Solaris 10 tweaks the 'lp' 
>> options, modifying simple "options" into "option=value".  As an 
>> example, if I specify "-o simplex" on the commandline, it is 
>> forwarded to the (Solaris 8) printer server as "simplex=true".
>>
>> Google, SunSolve, and docs.sun.com have been fruitless so far.
>>
>> [demime 1.01b removed an attachment of type text/x-vcard which had a 
>> name of Matthew_Stier.vcf]
>> _______________________________________________
>> sunmanagers mailing list
>> sunmanagers at sunmanagers.org
>> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Matthew.Stier at us.fujitsu.com  Tue Aug 19 17:57:50 2008
From: Matthew.Stier at us.fujitsu.com (Matthew Stier)
Date: Tue, 19 Aug 2008 17:57:50 -0400
Subject: SUMMARY [2]: Has anyone seen LP options changes under Solaris
	10 (05/08)?
In-Reply-To: <48A8F226.8060308@us.fujitsu.com>
References: <487BD028.5000705@us.fujitsu.com>
	<48A55029.7070101@uni-paderborn.de> <48A8F226.8060308@us.fujitsu.com>
Message-ID: <48AB41DE.9010301@us.fujitsu.com>

Okay Solaris 10 0508 (S10u5) users, this is what I have done to resolve 
this issue at my facility.  [Note: I am running my Unix print server on 
a Solaris 8 system.]

My solution was to write a script that took the list of options on the 
command-line, and converts specified key=value pairs, to  keys only.  I 
then tucked the script way in a directory already in PATH  
(/usr/spool/lp/bin) and then I made a one line change in each interface 
script to filter the options assignment through my new filter.

The script itself is simple.  It steps though each key=value pair on the 
command-line and after checking a few special conditions, it checks to 
see if the current key being checked, is on the list of keys to be 
converted.  If it is, and the value portion is "true", the key is added 
to the return string.  If the value portion is "false", "no" is prefixed 
to the key, and the revised key is added to the return string.  If it is 
neither a special condition, or on the list of keys to be converted, it 
is added, as-is.  When all the options are processed, the revised list 
of options is printed to standard out.

#!/usr/bin/sh
# Convert PAPI key=value pairs into boolean keys
# i=index, k=key, o=options, r=return
o="letter legal ledger 11x17 simplex duplex hduplex portrait landscape
banner 2up 4up manual manenv tray1 tray2 tray3 tray4 tray5 bin1 bin2
auto postscript pcl hpgl2 hpgl2_p A3 A4 A5 B4-JIS B5-JIS B5-ISO"
r=
for i in ${@}
do
        case "${i}" in
        job-sheets=standard )
                i="banner";;
        job-sheets=none )
                i="nobanner";;
        *)
                k=`expr "${i}" : "^\([^=]*\)=*"`
                for j in ${o}
                do
                        if [ "${i}" = "${j}=true" ]; then
                                i="${k}"
                                break
                        fi
                        if [ "${i}" = "${j}=false" ]; then
                                i="no${k}"
                                break
                        fi
                done;;
        esac
        r="${r:+${r} }${i}"
done
echo ${r}

To integrate this script into the interface script, I simply modify the 
assignment line, when the fifth parameter ($5) is assigned to a 
variable, so it is pre-processed by my script.

As an example, the netstandard script (which I use for the bases of all 
my custom interfaces scripts) assigns several positional parameters to 
variables, and shifts the parameters to leave only filenames on the 
commandline.  During these assignements, positional variable $5 is 
assigned to option_list.  ( option_list=$5 ) I use this assignment to 
implement the filter.  (option_list=`scriptname "${5}"`)

Note: At my facility, I have three classes of prints.  HP Laserjets, 
Xerox DocuCentre's and a Xerox plotter.  I use a different script for 
each class, but it is basically the same script for each class, with 
only the options to be converted, that is different.

Matthew Stier wrote:
> Found it, and the only fix is to modify the 'interfaces' scripts to 
> accept both pre and post PAPI boolean options.
>
> With Solaris 10 05/08, Sun implemented the PAPI interface, to support 
> multiple printing back ends.  (BSD/IPP/CUPS/...)
>
> The lp command calls the papiAttributeListFromString() function to 
> process each option. 
>
>     
> http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/print/bsd-sysv-commands/lp.c
>     http://src.opensolaris.org/source/s?defs=papiAttributeListFromString
>
> The papiAttributeListFromString calls a _parse_attribute_list() 
> function, who job it is to process the numerous option argument 
> formats.  Near the end of the function, it process boolean (key only) 
> options. The first thing the boolean key code does is convert the 
> boolean key into a key=value pair, with the key being set to a value of 
> "true".  It then test the first two characters of the key to see if it 
> begins with "no". If it does, it removes the "no" prefix, and sets the 
> new key to "false".  (ie: "banner" becomes "banner=true". "nobanner" 
> becomes "banner=false".)
>
>     http://src.opensolaris.org/source/s?refs=_parse_attribute_list
>     (lines 853-861)
>
> Thus the only fix is to go through each interfaces script, and modify 
> the code that process the options, to accept both 'key' and 'key=value' 
> options.
>
>
> Jan Dreyer wrote:
>   
>> Hi,
>>
>> though your message is quiet old, I would like to point out, that we 
>> encountered similar problems with clients on HP-UX. Our solution was 
>> to fix the scripts which send the data and make little tweaks to our 
>> own scripts in .../interfaces.
>>
>> Did you find another solution or a hint what SUN changed?
>>
>> Greetings
>> Jan Dreyer
>>
>>
>> Matthew Stier schrieb:
>>     
>>> I've just deployed Solaris 10 (05/08) to a number of users, and an 
>>> encountering a problem with some /etc/lp/interfaces scripts I wrote 
>>> years ago.  This problems has not been encountered on any previous 
>>> release of Solaris 10, or 9, 8, 7, ...
>>>
>>> It appears that the latest release of Solaris 10 tweaks the 'lp' 
>>> options, modifying simple "options" into "option=value".  As an 
>>> example, if I specify "-o simplex" on the commandline, it is 
>>> forwarded to the (Solaris 8) printer server as "simplex=true".
>>>
>>> Google, SunSolve, and docs.sun.com have been fruitless so far.
>>>
>>> [demime 1.01b removed an attachment of type text/x-vcard which had a 
>>> name of Matthew_Stier.vcf]
>>> _______________________________________________
>>> sunmanagers mailing list
>>> sunmanagers at sunmanagers.org
>>> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
>>>       
> _______________________________________________
> sunmanagers mailing list
> sunmanagers at sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Crist.Clark at globalstar.com  Fri Aug 22 20:53:54 2008
From: Crist.Clark at globalstar.com (Crist Clark)
Date: Fri, 22 Aug 2008 17:53:54 -0700
Subject: SUMMARY: Annoying Shell/Terminal Problem
References: <48AEDF23.8C45.0097.0@globalstar.com>
Message-ID: <48AEFD31.8C45.0097.0@globalstar.com>

Thanks to "francisco roque" for catching this one. He suggested
that a startup script was set to trap(1) INT. Yep. Some /etc/profile
and .profile scripts map INT to a null action with trap(1). Then
at the end of the script either (a) someone forgot to clear the
trap(1) or (b) someone changed it to ksh-specific syntax that sh
doesn't seem to understand (keep /etc/profile sh-clean people)
depending on the system.

Thanks.

On 8/22/2008 at 3:45 PM, "Crist Clark" <Crist.Clark at globalstar.com>
wrote:
> OK. I give up. I'm asking for help. We've got this annoying
> little problem with a number of machines. I'm not sure what
> this sysadmin does differently, but almost all of the boxes
> she builds show this behavior.
> 
> The primary annoyance is that Ctrl-C doesn't work like we
> want and expect. It doesn't actually interrupt a command.
> For example, if I,
> 
>   # tail -f /some/file
> 
> And then ^C to try to break out, it does not work. I can't
> break out. But if I ^Z, it does background it.
> 
> But ^C does do something. If I type,
> 
>   # echo howdy^Cecho doody
>   doody
> 
> The ^C canceled the first echo, but it doesn't cause the shell
> to kick back with a newline and prompt like I usually expect,
> 
>   # echo howdy^C
>   # echo doody
>   doody
> 
> That's the output on system with the glitch and one without for
> the exact same set of keystrokes.
> 
> As a plus, we have systems where the ^C will break out in the
> tail command example, but does still show the weird behavior
> when I run those echo commands from the shell.
> 
> I've looked at the shell environmental variables. I've looked
> at stty(1) settings. It doesn't look like the login shell,
> /sbin/sh and /bin/ksh both seem to do it. I'm stumped. Anyone
> seen this before?


BB<information contained in this e-mail message is confidential, intended
only for the use of the individual or entity named above. If the reader
of this e-mail is not the intended recipient, or the employee or agent
responsible to deliver it to the intended recipient, you are hereby
notified that any review, dissemination, distribution or copying of this
communication is strictly prohibited. If you have received this e-mail
in error, please contact postmaster at globalstar.com 
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Szabadkai.Eva at malev.hu  Wed Aug 27 05:11:54 2008
From: Szabadkai.Eva at malev.hu (=?iso-8859-1?Q?Szil=E1gyin=E9_Szabadkai_=C9va?=)
Date: Wed, 27 Aug 2008 11:11:54 +0200
Subject: SUMMARY: flarcreate question
Message-ID: <F56793DA4C266E4395316E9E4774B7760233AFDB@fhbmsg03.malev.hu>

The general consensus was that creating an archive from the root filesystem
couldn't be done without
using the exclude option of flarcreate (-x, -X)
It is not an  expected behaviour of flarcreate to archive only the root files
and not other filesystems by default.

Thanks to  Diego Brito Veiga, Turk Turco,  Gary Paveza, Kev Smith

Brgds, Eva


Original Question:

Hi Gurus,

how to create a flash archive only from the root filesystem and not from the
mounted filesystems,
without using exclude option?
I tried:
 #  flarcreate -n test /backup/flars/test.flar
and this command created a 100Gb file by archiving  all the mounted
filesystems.

Brgds, Eva
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From sunhux at gmail.com  Thu Aug 28 10:08:40 2008
From: sunhux at gmail.com (sunhux G)
Date: Thu, 28 Aug 2008 22:08:40 +0800
Subject: Summary: ufsdump from a remote server via ssh (Prng not seeded)
Message-ID: <60f08e700808280708u223fa271yb22aeeff769a652b@mail.gmail.com>

Thanks to various replies, but I'll just quote Julian (who emailed
me the patch), Duncan  &  Ric  to minimize overlapping answers:

=========================================

You should install sun patch 112438-02 which is the patch that
installs /dev/random and that will seed openssh.  There is also
a shell script that can install the patch live (no reboot).

There may be a newer version of that patch



With your ufsdump across the network, if you're still having
difficulties with sshd and PRNG (sounds like you need the
/dev/random patch from Sun, google for this and entropy), then a
simpler alternative for transporting the data is to get and install
netcat (from sunfreeware.com) on both machines.

Then, instead of something like:

ON HOST1: ufsdump 0uf - test.txt | ssh HOST2 "gzip >/tmp/host1.dmp.gz"

You'd first tell HOST2 to listen on port 10000 (your choice) and
capture what it gets:

nc -l -p 10000 | gzip > /tmp/host1.dmp.gz

then on HOST1, you'd generate the data stream and get netcat to send it
to the listening port on HOST2:

ufsdump 0f - | nc HOST2 10000

10000 is an arbitrary port you choose that is:
a). not used on host2,
b). not firewalled when HOST1 tries to connect to it on HOST2.

You may need to CTRL-C HOST1's ufsdump | nc command when the ufsdump
has finished, for reasons never very clear to me netcat often hangs around.
However, as ufsdump itself reports when it's finished, and you can also
use another connection to HOST2 to monitor the size of /tmp/host1.dump.gz,
or indeed the cpu usage of gzip, this should be ok.


=========================================

Q: I was able to ssh into this Solaris 8 server but from this
    Solaris 8 server, I'm not able to ssh out to other server
A: Hmm, someone installed some ssh package before you
    got there.  Odd that they would install the server and
    not the client though.  Look in /etc/rc?.d and see
    what script starts sshd and where it's config file is.





Thanks


On Sat, Aug 16, 2008 at 8:07 PM, sunhux G <sunhux at gmail.com> wrote:

> Thanks, not a path issue but Openssh is not there in Solaris 8.
>
> The command is :
> ufsdump 0uf - test.txt | ssh slagpo at 10.51.101.5 "cd /tmp; cat > test.dmp;
> gzip test.dmp"
> (tested between 2 Solaris 10 servers)
>
> (One reply corrected my original command :
>   # nohup ufsdump 0uf - /  | ssh myuserid at 10.51.101.5 "cd /ora/destdir;
>     zip ufsdump.zip -"
>
>
> I'll need to install from www.sunfreeware.com  Openssh (& the
> pre-requisites Openssl & lastly libgcc) but after restarting (by issuing
> "/usr/sbin/sshd2 restart" ), ssh now has a message :
>
> # /usr/local/bin/ssh myuserid at 10.51.101.5 <myuserid at 10.51.101.5>
> PRNG is not seeded
>
> (Some of those suggestions from google did not resolve PRNG error
>   above)
>
> NFS share is not permitted for security reason.
>
>
> Thanks again,
> U
>
> On 8/15/08, sunhux G <sunhux at gmail.com> wrote:
>>
>> Hi
>>
>> I'm exploring a solution to do "ufsdump" from a Solaris 8 server (which
>> currently also has an 'ssh command not found problem') to a remote
>> Solaris 10 server (which has a lot of disk space) :
>>
>> On Solaris 8 server :
>> # ssh
>> ksh: ssh:  not found
>>
>>
>> The command below on Solaris 8 did not work (partly the
>> ssh problem above & possibly I'm missing something) :
>>
>> # nohup ufsdump 0uf - /  | ssh myuserid at 10.51.101.5 "cd /ora/destdir; zip
>> -" &
>>
>> where /ora/destdir  is permitted for writing by the remote user myuserid.
>>
>>
>> What's the missing bits?
>>
>>
>> Also, suppose instead of using the remote Solaris 10 server, can I
>> substitute
>> it with a Windows ssh server (which also has a lot of disk space)?
>>  Problem is Windows XP supports max file size of 4Gb only if I'm not wrong
>>
>>
>> TIA
>> U
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From Szabadkai.Eva at malev.hu  Thu Aug 28 04:25:01 2008
From: Szabadkai.Eva at malev.hu (=?iso-8859-1?Q?Szil=E1gyin=E9_Szabadkai_=C9va?=)
Date: Thu, 28 Aug 2008 10:25:01 +0200
Subject: FW: SUMMARY: flarcreate question
Message-ID: <F56793DA4C266E4395316E9E4774B7760218A02E@fhbmsg03.malev.hu>

Thanks to Tim Bradshaw who suggested another way to create a root flash
archive.

On 27 Aug 2008, at 17:33, Tim Bradshaw wrote:

> A good trick is to use fssnap to create a snapshot of /, and then mount that
and use flarcreate -R ... to create a flar of it.

Brgds, Eva
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From bj+sunmanagers at glue.ch  Thu Aug 28 12:36:21 2008
From: bj+sunmanagers at glue.ch (Beat Jucker)
Date: Thu, 28 Aug 2008 18:36:21 +0200
Subject: SUMMARY: GRUB: booting from cloned root disk (how to)
Message-ID: <20080828183621.B18600@cicero.glue.ch>

Thanks to Doug Yatcilla. He gave me the missing puzzle:

  mount /dev/dsk/c1t1d0s0 /mnt
  bootadm update-archive -R /mnt     

-- Beat

----------------------------------------------------------------------

> Well, this is how I see things;
>
> c1t0d0s0 is the disk with your original Solaris 10 system.  That disk
> is hd(0,0,a) from GRUB's point of view.  The c1t0d0s0 should have
> "/pci.../sd at 0,0,a" specified in the bootenv.rc file.
>
> If your disk controller will only boot from c1t0d0s0, then the GRUB
> bootblocks are read from that disk as are all the files in /boot/grub
> (like menu.lst).  So, that disk needs to be present when the system  
> boots.
> 
> But, you may specify a different disk from which Solaris will be
> booted.  I would verify that using "root (hd0,0,a)" in GRUB will boot
> Solaris from c1t0d0s0.
> 
> Next, mount c1t1d0s0 on /mnt and verify that /mnt/etc/vfstab
> has the root filesystem specified as c1t1d0s0.  Also, verify that
> /mnt/boot/solaris/bootenv.rc has bootpath set to "/pci.../sd at 1,0,a"
> 
> Next, mount c1t1d0s0 on /mnt and verify that /mnt/etc/vfstab
> has the root filesystem specified as c1t1d0s0.  Also, verify that
> /mnt/boot/solaris/bootenv.rc has bootpath set to "/pci.../sd at 1,0,a"
> 
> Then, reboot the system and specify "root (hd1,0,a)" in GRUB to boot
> Solaris from c1t1d0s0.
> 
> If that doesn't work, then I am stumped.
> 
> If it doesn't work, I would use Live Upgrade (lucreate & luactivate
> commands) to make a copy of your current boot environment from     
> c1t0d0s0 to c1t1d0s0.  This will automatically do exactly what you are
> trying to manually.  LU will automatically adjust GRUB, bootenv.rc,   
> /etc/vfstab, etc.  It is probably the safest thing to do.
> 
> Good luck,
> Doug 

Hello Beat,

I forgot about the bootadm command.  I think you need to run
"bootadm -R /mnt update-archive" in order for the changes you made to
/boot/solaris/bootenv.rc to be included in the ramdisk image of
Solaris that is booted by GRUB.

More info here::
http://www.sun.com/bigadmin/features/articles/grub_boot_solaris.jsp
http://flux.org.uk/howto/solaris/fix_boot_archive

The lesson is that the booting process for x86 Solaris is more/overly
complicated and prone to errors compared to SPARC.  Live Upgrade     
simplifies things, though.

-Doug
     
Hello SUN sysadmins

We have a Sun Fire x4240 (fully patched Solaris 10).
Until now I have worked only with Sparc & Solaris9. 
I'm not so familiar with x86 & Solaris10. 

I want to be able booting from a cloned disk.
Because ZFS isn't ready for system FS I have cloned the system disk
manually (just a small modified script already working for our     
Solaris9 Sun Fires):

  - partition the clone disk, make it bootable
    fdisk -B /dev/rdsk/c1t1d0p0
    prtvtoc /dev/rdsk/c1t0d0s2 | fmthard -s - /dev/rdsk/c1t1d0s2
    installgrub -fm /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c1t1d0s0
    
  - create filesystem and copy all files to the clone disk (using snapshot
    and ufsdump, adjusting cloned /etc/vfstab afterwards)
    
  - enhance GRUB boot menu to hd1 (default: root=hd0,0,a)
    title Solaris 10 clone disk
    root (hd1,0,a)
    kernel /platform/i86pc/multiboot
    module /platform/i86pc/boot_archive
    
So far everything seems OK but when I boot the system (GRUB boot menu
"Solaris 10 clone disk") it boots always from default root (c1t0d0) and
not from cloned disk (c1t1d0). Even when I change/edit the boot disk manually
from GRUP boot menu: root (hd0,0,a) --> root (hd1,0,a) it boots from c1t0d0 ...

Did I miss something?

Because I'll use zfs raid feature for user data I didnt touch the builtin
Raid controller ASR-5805. But after checking disk status with the command
"arcconf getstat 1 al" I can see only disk-0 has "Bootable=Yes" and all  
other disks have "Bootable=No". Could this be the problem?
How to set disk-1 as "Bootable=Yes" (couldn't find this task in the manual)?

Best regards
-- Beat     
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From jacob.ritorto at gmail.com  Fri Aug 29 15:03:46 2008
From: jacob.ritorto at gmail.com (Jacob Ritorto)
Date: Fri, 29 Aug 2008 15:03:46 -0400
Subject: Summary: rbac cluebyfour?
Message-ID: <1f3f8f1d0808291203u2b71547x6c51b0e8c7cf3c2c@mail.gmail.com>

Thanks to all who responded.  Some actually answered my question and
gave instructions on how to do it with big crazy RBAC, while others
admonished me to not use such a big hammer when setfacl (POSIX file
access control list settings) should to the trick neatly.  This is
Solaris 8 SPARC 7/01, by the way..

I chose the facls method.  It was a daunting mess due to missing
documentation on Docs.Sun.COM.  Fortunately Softpanorama came through
as usual with the missing info:
http://www.softpanorama.org/Solaris/ACL/index.shtml

To give a bit more detail on what my scenario, we have a
multi-customer host that dishes out confidential, unique info to
clients.  Obviously it;s necessary that they not see each others'
files, but the user oracle had to be able to dish out said files.  So
'trivial' or traditional unix perms were not appropriate to the task.

Solution:

1) Create individual home dirs for users that have 700 traditional
unix permissions.  This way nobody can peer at others' files.

2) set initial default inheritance capability in facls (the part they
failed to mention on docs.sun.com).  This enables you to specify the
acl behaviour of files dropped in the directory:

setfacl -r -m default:user::rwx,default:group::---,default:other:---,default:mask:rwx
~bob/

3) give oracle and the customer the ability to manipulate the files,
despite not being listed in trivial acl:
setfacl -r -m default:user:oracle:rwx ~bob/
setfacl -r -m default:user:bob:rwx ~bob/

4) repeat similar for other clients

5) check it and confirm with real world testing:
root at Lumpy~bob[8]14:22#getfacl ~bob/
# file: bob
# owner: bob
# group: clients
user::rwx
user:oracle:rwx         #effective:rwx
group::---              #effective:---
mask:rwx
other:---
default:user::rwx
default:user:oracle:rwx
default:user:bob:rwx
default:group::---
default:mask:rwx
default:other:---


Seems good now.

thx
jake



I wrote:
>         Is it appropriate to use RBAC to allow one user to write files
> in the directory of another user?  We need 600 perms for user bob's
> home but want user oracle to dump output there.  If RBAC is the right
> way to do this, would someone give recipe/example please?  I can't
> digest the entire RBAC documentation before deadline.
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers

From anepomn at gmail.com  Fri Aug 29 18:53:50 2008
From: anepomn at gmail.com (Aleksandr Nepomnyashchiy)
Date: Fri, 29 Aug 2008 18:53:50 -0400
Subject: SUMMARY: Command doesn't work in cron but works in shell
Message-ID: <26924da90808291553k7eed264aq2870be64a2ef4c0c@mail.gmail.com>

Many thanks to everyone who replied and especially to Ric Anderson,
Bill R. Williams , Francisco Roque and Michael Maciolek who pointed me
to the following fragment of the man page:
 "The sixth field of a line in a crontab file is a string that
 is executed by the shell at the specified times. A percent
 character in this field (unless escaped by \ ) is translated
 to a NEWLINE character"

The following cron line has been successfully tested:
10 23 * * * mv /path/to/my/file.txt  /path/to/my/file.txt`date +\%Y\%m\%e`


=========== Original question ==============

The following command works in shell :
root at host ~ # mv /path/to/my/file.txt  /path/to/my/file.txt`date +%Y%m%e`



But doesn't work in cron :
# rotate my log
10 23 * * * mv /path/to/my/file.txt  /path/to/my/file.txt`date +%Y%m%e`

Root is getting an email :
>From root Wed Aug 27 23:10:01 2008
Return-Path: <root>
Received: (from root at localhost)
To: root
Subject: Output from "cron" command


Your "cron" job on host
mv /path/to/my/file.txt  /path/to/my/file.txt`date +

produced the following output:

mv: /path/to/my/file.txt and /path/to/my/file.txt are identical


Looks like  `date +%Y%m%e`  doesn't get interpreted the way I expect
when running from cron.


Thank you,
Aleksandr
_______________________________________________
sunmanagers mailing list
sunmanagers at sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers