Bill Bradford | 14 Mar 05:33 2009
Picon

ADMIN NOTE: Server upgrade soon

Hey folks,

A status update on the machine that hosts all these lists, SunHELP, an IRC
server, and lots of folks' DNS, email handling, and shell accounts:

The long-delayed upgrade of ohno.mrbill.net from an E420R (quad 450s, 4G
RAM, dual 18G hard drives) to a T1000 (8 cores, 32 threads, 8G RAM) will
finally happen in the next 30-60 days.  I'll be going to S10u6 with ZFS
root of course.

I've taken up a collection from some of the major users of the system
(shell, email, etc) and I'll be upgrading the 80G drive that came with the
system to a 1T drive (most likely a Western Digital "Caviar Black"
WD1001FALS, unless someone has other suggestions).  

I thought about putting in one of the WD10EADS "Caviar Green" drives but
they've only got a 3-year warranty and I don't think they're really meant
for heavy-duty server use.  I'd love to put the WD "RE3" Raid Edition drive
in, but they're $50 more than the WD1001FALS.

Depending on funds, I will probably also purchase another WD10EADS drive to
put in my rsnapshot backup server here at home, and backup *all* of the
user homedirs and data on the new machine and eliminate the " you're
responsible for your own backups" policy.  

If any member of these lists has gotten use and enjoyment out of the
services I provide (12 years and counting so far!) and would like to
contribute to the upgrade fund, please let me know offlist.  

There will be a few hours of downtime over a weekend sometime in the next
(Continue reading)

Bill Bradford | 17 Mar 20:00 2009
Picon

Re: [geeks] ADMIN NOTE: Server upgrade soon

Just wanted to say a very big thank to you to everyone that has
contributed so far.  Looks like I'll be getting that "Raid Edition" drive
after all!

Bill

On Fri, Mar 13, 2009 at 11:33:06PM -0500, Bill Bradford wrote:
> Hey folks,
> A status update on the machine that hosts all these lists, SunHELP, an IRC
> server, and lots of folks' DNS, email handling, and shell accounts:
> The long-delayed upgrade of ohno.mrbill.net from an E420R (quad 450s, 4G
> RAM, dual 18G hard drives) to a T1000 (8 cores, 32 threads, 8G RAM) will
> finally happen in the next 30-60 days.  I'll be going to S10u6 with ZFS
> root of course.
> I've taken up a collection from some of the major users of the system
> (shell, email, etc) and I'll be upgrading the 80G drive that came with the
> system to a 1T drive (most likely a Western Digital "Caviar Black"
> WD1001FALS, unless someone has other suggestions).  
> I thought about putting in one of the WD10EADS "Caviar Green" drives but
> they've only got a 3-year warranty and I don't think they're really meant
> for heavy-duty server use.  I'd love to put the WD "RE3" Raid Edition drive
> in, but they're $50 more than the WD1001FALS.
> Depending on funds, I will probably also purchase another WD10EADS drive to
> put in my rsnapshot backup server here at home, and backup *all* of the
> user homedirs and data on the new machine and eliminate the " you're
> responsible for your own backups" policy.  
> If any member of these lists has gotten use and enjoyment out of the
> services I provide (12 years and counting so far!) and would like to
> contribute to the upgrade fund, please let me know offlist.  
> There will be a few hours of downtime over a weekend sometime in the next
(Continue reading)

Shannon Hendrix | 19 Mar 19:02 2009

latest system monitoring for Solaris 10 x86

Does anyone know if Solaris 10 has any available updates or third  
party packages to monitor system sensors on an X86 machine?

In my case I'd like a few simple things like temperature for CPU and  
drives for a Dell T105 server (which runs Solaris 10 very well BTW).

--

-- 
"Where some they sell their dreams for small desires."
_______________________________________________
SunHELP maillist  -  SunHELP <at> sunhelp.org
http://www.sunhelp.org/mailman/listinfo/sunhelp

Bill Bradford | 25 Mar 00:08 2009
Picon

ADMIN NOTE: Thank you!

Just another admin note.

Thank you, everyone.  Thanks so much for your contributions to the server
upgrade fund.  I never expected the generous response I got to my upgrade
status email.

Due to the contributions from everyone, I've got the new WD RE3 (Raid
Edition) 1T drive for the new system sitting here on my desk, along a pair
of WD "green" 1T WD10EADS drives that will be ZFS mirrored and used for the
local rsnapshot backups of the main system (which is located in Austin, I'm
in Houston).  

I've also got a third WD 1T drive in an external USB enclosure that will
be used for monthly "offsite offsite" backups; I'll be rsyncing the
rsnapshots to it and then keeping it in my office at work.

I've received plenty of funds and have everything I need for the ugprade
process; no more monetary contributions are needed (unless you want it to
go towards the "buy Ms.  Bradford a nice dinner for putting up with
hardware all over the dining room table" fund).  

We now return you to normal "If you like the services I provide, consider
buying something off http://wishlist.sunhelp.org" programming.

I'll be building the Jumpstart server this coming weekend, and either
shipping it and the drive to OnRamp,  or going with my wife to a doctor
appointment in Austin and doing the work in person, around the 31st.  If
all goes well, you can expect the migration to happen in mid-April.

Again, thank you everyone.  The new server should be awesome.
(Continue reading)

Michael Karl | 30 Mar 10:35 2009
Picon

Sol10 - fsck problem / on SunSparc-AXMP

Hello,

some days ago I saw some memeory-errors (on module u0701) on my 
SunSparc-AXmp (last recommended patch-cluster two years ago)

Yesterday I replaced this module.

After reboot there are two free inodes in / and I should run fsck -o f 
/dev/md/rdsk/d0. I done this, the superblock was fixed, but after reboot 
I got the same error. I could see via google, that some people had the 
same problem ... but I found no solution.

Is the option "fsck -o f" correct?

What's the best solution to solve this issue ?

Thank in advance

Michael
_______________________________________________
SunHELP maillist  -  SunHELP <at> sunhelp.org
http://www.sunhelp.org/mailman/listinfo/sunhelp

Michael Karl | 30 Mar 13:36 2009
Picon

Re: [50% solved, but need help ; -)]Sol10 - fsck problem / on SunSparc-AXMP

Hello,

sorry, I made a mistake ;-(

There was a mirror on root ... yesterday I metadetached the mirror (I'm 
thinking, that's my mistake) ...  and then the reboot starts from the 
metadetached mirror by default from openbootprom ... I don't remember, 
who had done this openbootprom-envmany years ago.

I corrected this circumstance and now the machine is starting well.

But now my question:
This Sun is more then 8 years old (4x 360Mhz UltraII, 2GB RAM). The 
Solaris 10 has no atctual recommended patch-cluster.
What is the best und easiest way to install the patch-cluster on the 
first disk and have the option of the now working Solaris 10 on the 
second metadetached mirror-disk?

Regards Michael

 
Michael Karl schrieb:
> Hello,
>
> some days ago I saw some memeory-errors (on module u0701) on my 
> SunSparc-AXmp (last recommended patch-cluster two years ago)
>
> Yesterday I replaced this module.
>
> After reboot there are two free inodes in / and I should run fsck -o f 
(Continue reading)

Sandwich Maker | 30 Mar 15:24 2009
Picon

Re: [50% solved, but need help ; -)]Sol10 - fsck problem / on SunSparc-AXMP

" From: Michael Karl <mk <at> lexcom-net.de>
" 
" 
" This Sun is more then 8 years old (4x 360Mhz UltraII, 2GB RAM). The 
" Solaris 10 has no actual recommended patch-cluster.
" What is the best und easiest way to install the patch-cluster on the 
" first disk and have the option of the now working Solaris 10 on the 
" second metadetached mirror-disk?

probably your best approach is to un-mirror the disks, so that you can
boot from either independently.  re-mirror when you decide which one
you want to go with.
________________________________________________________________________
Andrew Hay                                  the genius nature
internet rambler                            is to see what all have seen
adh <at> an.bradford.ma.us                       and think what none thought
_______________________________________________
SunHELP maillist  -  SunHELP <at> sunhelp.org
http://www.sunhelp.org/mailman/listinfo/sunhelp

Michael Karl | 30 Mar 17:14 2009
Picon

Re: [50% solved, but need help ; -)]Sol10 - fsck problem / on SunSparc-AXMP

Hello Andrew,

Sandwich Maker schrieb:
> " From: Michael Karl <mk <at> lexcom-net.de>
> " 
> " 
> " This Sun is more then 8 years old (4x 360Mhz UltraII, 2GB RAM). The 
> " Solaris 10 has no actual recommended patch-cluster.
> " What is the best und easiest way to install the patch-cluster on the 
> " first disk and have the option of the now working Solaris 10 on the 
> " second metadetached mirror-disk?
>
> probably your best approach is to un-mirror the disks, so that you can
> boot from either independently.  re-mirror when you decide which one
> you want to go with.
>   
That was my first idea ... but I'm making something wrong, that the 
metadetached ROOT is damaged. Is in Sol10 something different to Sol8/9 ?

Sorry about my simple question ... I had done this so many times years 
ago ... but I had an accident  2,5 years ago ... after 4 operations I'm 
now not fit at all.

Thank you for your help

Michael
> ________________________________________________________________________
> Andrew Hay                                  the genius nature
> internet rambler                            is to see what all have seen
> adh <at> an.bradford.ma.us                       and think what none thought
(Continue reading)

listmail | 30 Mar 17:53 2009
Picon

Re: [50% solved, but need help ; -)]Sol10 - fsck problem / on SunSparc-AXMP

metastat says there are problems?  Did you run metaroot to revert vfstab 
to use the physical slice and then metadetach all (secondary) sub 
mirror(s) before rebooting?  And after you rebooted did you use 
metaclear on the (secondary) sub mirrors and then metaclear -r each 
toplevel mirror?  And then perhaps also use metadb to delete the state 
replicas so that metadb -i returns nothing.   You can still boot right?

please show
# metastat

# metastat -p

# metadb -i

# cat /etc/vfstab

Michael Karl wrote:
> Hello Andrew,
>
> Sandwich Maker schrieb:
>> " From: Michael Karl <mk <at> lexcom-net.de>
>> " " " This Sun is more then 8 years old (4x 360Mhz UltraII, 2GB RAM). 
>> The " Solaris 10 has no actual recommended patch-cluster.
>> " What is the best und easiest way to install the patch-cluster on 
>> the " first disk and have the option of the now working Solaris 10 on 
>> the " second metadetached mirror-disk?
>>
>> probably your best approach is to un-mirror the disks, so that you can
>> boot from either independently.  re-mirror when you decide which one
>> you want to go with.
(Continue reading)

Michael Karl | 30 Mar 18:32 2009
Picon

Re: [50% solved, but need help ; -)]Sol10 - fsck problem / on SunSparc-AXMP

Hi,

listmail schrieb:
> metastat says there are problems?  
everything was OK before metadetach the submirrors.
> Did you run metaroot to revert vfstab to use the physical slice 
no ... why ... if one hd fail, it should work without using metaroot ... 
I'm thinking.
> and then metadetach all (secondary) sub mirror(s) before rebooting?
yes
> And after you rebooted did you use metaclear on the (secondary) sub 
> mirrors and then metaclear -r each toplevel mirror?
no ... why ?
>   And then perhaps also use metadb to delete the state replicas so 
> that metadb -i returns nothing.   
ok
> You can still boot right?
yes ... with the first disk after I corrected the 
openbootprom-enviroment. But this misconfiguration of the obp was the 
reason, that the metadetached mirror-disk didn't start with 
inode-errors. I'm thinking, that my way was wrong yesterday.
1. metadetach all submirrors
2. init 5
3. replacing ram-module
4. switch on and boot without any parameters
>
> please show
> # metastat
Sorry ... german output ;-)

(Continue reading)


Gmane