Re: [e-smith-devinfo] RAID 1

2001-09-14 Thread Dan York

[EMAIL PROTECTED] wrote:
 
 In case of a drive failure in e-smith software raid 1 what would be the
 procedure to restore the system back to normal.  Anybody has a How to?

I have a HOWTO in the queue on precisely this subject.  I have been
working on getting V5 docs out the door, and post-V5 am planning to
attack the queue of things that need HOWTOs but since the question
has been asked, I'll see what I can do to get this one out sooner
versus later.

Regards,
Dan

-- 
Dan York, Director of Training[EMAIL PROTECTED]
Ph: +1-613-751-4401 Cell: +1-613-263-4312 Fax: +1-613-564-7739 
Mitel Network Corporation Network Server Solutions Group 
150 Metcalfe St., Suite 1500, Ottawa,ON K2P 1P1 Canada
http://www.e-smith.com/http://www.mitel.com/   

--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




RE: [e-smith-devinfo] RAID 1

2001-09-14 Thread Trevor Ouellette

I will try it out this weekend.

Trev.

 -Original Message-
 From: Darrell May [mailto:[EMAIL PROTECTED]]
 Sent: Thursday, September 13, 2001 6:30 PM
 To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
 Subject: Re: [e-smith-devinfo] RAID 1



 [EMAIL PROTECTED] said:

  In case of a drive failure in e-smith software raid 1 what would be the
  procedure to restore the system back to normal.  Anybody has a How to?

 Found this document:

 http://www.linuxdoc.org/HOWTO/Software-RAID-HOWTO.html

 Which states in Chapter 6:
 --
 - Power down the system
 - Replace the failed disk
 - Power up the system once again.
 - Use raidhotadd /dev/mdX /dev/sdX to re-insert the disk in the array

 Have coffee while you watch the automatic reconstruction running and
 that's it.
 --

 I believe /proc/mdstat may show the activity.  I've not tried this
 myself.  If anyone has a test system, using software RAID ready to go,
 how about giving this a try and reporting back your results :)

 Regards,

 --
 Darrell May
 DMC NETSOURCED.COM
 http://netsourced.com

 --
 Please report bugs to [EMAIL PROTECTED]
 Please mail [EMAIL PROTECTED] (only) to discuss security issues
 Support for registered customers and partners to [EMAIL PROTECTED]
 To unsubscribe, e-mail: [EMAIL PROTECTED]
 For additional commands, e-mail: [EMAIL PROTECTED]
 Archives by mail and
http://www.mail-archive.com/devinfo%40lists.e-smith.org



--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1

2001-09-13 Thread Thomas E. Keiser

I've asked this a couple of times; even had a Mitec sales engineer promising to
e-mail me on this subject, but there doesn't appear to be a simple answer. The
suggestion that made the most sense, was to get a 3ware 6200 hardware raid card
(about $125) use that instead of the software raid. The card has built-in bios
recovery, and is well enough designed to give a good speed improvement over
either software raid or a single drive.

Hope this helps,

Tom

[EMAIL PROTECTED] wrote:

 In case of a drive failure in e-smith software raid 1 what would be the
 procedure to restore the system back to normal.  Anybody has a How to?

 --
 Please report bugs to [EMAIL PROTECTED]
 Please mail [EMAIL PROTECTED] (only) to discuss security issues
 Support for registered customers and partners to [EMAIL PROTECTED]
 To unsubscribe, e-mail: [EMAIL PROTECTED]
 For additional commands, e-mail: [EMAIL PROTECTED]
 Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org


--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1

2001-09-13 Thread Darrell May


[EMAIL PROTECTED] said:

 In case of a drive failure in e-smith software raid 1 what would be the 
 procedure to restore the system back to normal.  Anybody has a How to?

Found this document:

http://www.linuxdoc.org/HOWTO/Software-RAID-HOWTO.html

Which states in Chapter 6:
--
- Power down the system 
- Replace the failed disk 
- Power up the system once again. 
- Use raidhotadd /dev/mdX /dev/sdX to re-insert the disk in the array 

Have coffee while you watch the automatic reconstruction running and 
that's it. 
--

I believe /proc/mdstat may show the activity.  I've not tried this 
myself.  If anyone has a test system, using software RAID ready to go, 
how about giving this a try and reporting back your results :)

Regards,

-- 
Darrell May
DMC NETSOURCED.COM
http://netsourced.com

--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1

2001-09-13 Thread Richard Ford

The high point is software anyway as well.

Like winmodems are software modems - but still have hardware.

Go for the 3ware or a compaq or HP option.

Richard.


- Original Message -
From: Thomas E. Keiser [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Cc: [EMAIL PROTECTED]
Sent: Friday, September 14, 2001 8:30 AM
Subject: Re: [e-smith-devinfo] RAID 1


 I've asked this a couple of times; even had a Mitec sales engineer
promising to
 e-mail me on this subject, but there doesn't appear to be a simple answer.
The
 suggestion that made the most sense, was to get a 3ware 6200 hardware raid
card
 (about $125) use that instead of the software raid. The card has built-in
bios
 recovery, and is well enough designed to give a good speed improvement
over
 either software raid or a single drive.

 Hope this helps,

 Tom

 [EMAIL PROTECTED] wrote:

  In case of a drive failure in e-smith software raid 1 what would be the
  procedure to restore the system back to normal.  Anybody has a How to?
 
  --
  Please report bugs to [EMAIL PROTECTED]
  Please mail [EMAIL PROTECTED] (only) to discuss security issues
  Support for registered customers and partners to [EMAIL PROTECTED]
  To unsubscribe, e-mail: [EMAIL PROTECTED]
  For additional commands, e-mail: [EMAIL PROTECTED]
  Archives by mail and
http://www.mail-archive.com/devinfo%40lists.e-smith.org


 --
 Please report bugs to [EMAIL PROTECTED]
 Please mail [EMAIL PROTECTED] (only) to discuss security issues
 Support for registered customers and partners to [EMAIL PROTECTED]
 To unsubscribe, e-mail: [EMAIL PROTECTED]
 For additional commands, e-mail: [EMAIL PROTECTED]
 Archives by mail and
http://www.mail-archive.com/devinfo%40lists.e-smith.org




--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1 installation problem

2001-07-15 Thread Des Dougan

At 20:14 10/7/2001, Des Dougan wrote:
I will have to annoy the family at the weekend and take the server offline...

Oh my... Never a truer word spoken. I've basically been down 24 hours, all 
in, but am finally up again with one minor problem (of which more later).

I did a backup to desktop (no tape drive yet). This took several hours (1.2 
GB), but finally completed OK. I then brought down the server and dug into 
the diagnostics and EISA configuration options (the box is an early Dell 
PowerEdge server with an EISA bus and SCSI interfaces). The option to 
low-level format is part of the EISA utilities. I ran it on disk 0, and it 
ran for well over an hour, finally ending with a Caution message - 
apparently caused by a time-out. I then ran disk 1, and it ran about an 
hour and ended with the same message. Re-booted, and got an error message 
in the SCSI BIOS about head/cyl. translation. I eventually worked out that 
the DOS  1024 cylinder setting was on and was affecting the geometry. 
Re-set, went back and re-formatted again. This time disk 1 failed with a 
medium error. Retried - same result. Rebooting showed the drive in the 
BIOS, but it couldn't be seen by either Linux or DOS. I installed e-smith 
(I actually used the RAID 1 setting again - it installed, but used only 
/dev/sda, of course). In my attempts to try to figure out if this was a 
hardware problem or simply a configuration issue, I did the low-level 
format again, and somehow managed to blow away /dev/sda too. As it was now 
well after 1.00 a.m., I left it till this morning. Fortunately my test 
server has a SCSI disk (also a Seagate Barracuda), so I swapped it in, and 
it booted OK (phew!!).

I then ran a clean install on that disk, configured it, and set the restore 
from desktop going. This worked fine, as far as I could see - could see and 
access shares, that is, until I tried to log in to the Manager - it 
wouldn't accept my password. Swapped the video back across and tried from 
the console - nada for both root and admin (same password, of course). I 
was positive I had not made any error with passwords (I used the same as 
the previous one on the rebuild), but was basically stuck (I'd also changed 
admin to log in, although that wouldn't have made much difference).

I tried a few things, including checking that passwd and shadow had been 
included in the backup file; booted using tomsrtbt (couldn't mount 
/dev/sda6); and ran the update option. I eventually re-installed a clean 
system and re-restored, this time logging in to the console at the earliest 
opportunity. This was fortuitous, as the same problem occurred again. It 
seems that the /etc/shadow file had been zapped - it was there but was 0 
bytes. I tried copying /etc/shadow- to the main file, but this didn't help 
either. I manually set the root password before doing the final reboot 
following the restore, so that I was able to log in after the reboot. I had 
to also reset the admin passwd to access the Manager.

As I noted, everything seems OK except that I can't get in using ssh. I 
hadn't re-applied the 4.1.2 updates before doing the restore, and got this 
message in the log:

Jul 15 18:00:16 jeeves sshd: RSA key generation failed
Jul 15 18:00:16 jeeves e-smith-bg: Generating SSH2 RSA host key: [ FAILED ]^M

I applied the 3 RPMs in the update directory and re-tried the ssh settings, 
but get the same message. I would appreciate any help on resolving this 
issue. I will also report the passwd/shadow problem to the bugs@ account - 
after the second iteration, I am more than convinced this is a bug of some 
sort in the restore routines.





Des Dougan


--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1 installation problem

2001-07-10 Thread Des Dougan

Dan,

Many thanks for your reply. Umm (embarrassed look) yes, Raid 1 is what I 
installed - standard e-smith settings. The two disks are both Seagate 
drives, both 4 GB, but have different P/Ns. I will have to annoy the family 
at the weekend and take the server offline...

At 06:53 10/7/2001, Dan York wrote:
Just to be sure we are clear... we're talking about RAID 1, right?



Des


--
Please report bugs to [EMAIL PROTECTED]
Please mail [EMAIL PROTECTED] (only) to discuss security issues
Support for registered customers and partners to [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Archives by mail and http://www.mail-archive.com/devinfo%40lists.e-smith.org




Re: [e-smith-devinfo] RAID 1 Status

2001-04-26 Thread Gordon Rowell

On Mon, Dec 11, 2000 at 08:57:10PM +0200, Brandon Friedman 
[EMAIL PROTECTED] wrote:
 Is there way to check the status of the raid partition?
 
 How do we know it is really working? Other than physical disconnecting one
 drive?
 
 If one drive crashes , how do you reinitialise the mirror once you have
 installed the new drive?

These are all good questions which need to go into the manual once
it's written :-), this is a developer preview. /usr/doc/raidtools-0.90
would be a a good place to start and/or searching for Linux RAID1 on
your favourite Linux search engine.

 Will there be inclusion of RAID 0 

No, RAID0 is just plain unsafe. If a disk crashes you have junk as far
as the head can scan :-) With RAID1 you can use one of the mirror halves.

RAID0+1 (sometimes called RAID10) may be a possibility in the future.

 or 5 ???

No. RAID5 is a performance pig and disk is cheap. If you need a disk
farm, buy a box that does RAID5 in hardware. That way you have a
cache between the host and the disks, so the read-modify-write cycles
happen in the box, not while your system is waiting for them.

And for completeness, that's a no to RAID3 and RAID4.

Gordon
--
  Gordon Rowell [EMAIL PROTECTED]
  http://www.e-smith.org (development)  http://www.e-smith.com (corporate)
  Phone: +1 (613) 564 8000 ext. 4378Fax: +1 (613) 564 7739
  e-smith, inc. 1500-150 Metcalfe St, Ottawa, ON K2P 1P1 Canada



Re: [e-smith-devinfo] RAID 1 Status

2001-04-26 Thread Peter Green

* Gordon Rowell [EMAIL PROTECTED] [001211 14:15]:
  Will there be inclusion of RAID 0 
 
 No, RAID0 is just plain unsafe. If a disk crashes you have junk as far
 as the head can scan :-) With RAID1 you can use one of the mirror halves.
 
 RAID0+1 (sometimes called RAID10) may be a possibility in the future.

Just a minor nit to pick here, and a suggestion. RAID10 probably refers more
often to RAID1+0, rather than RAID0+1. Also, if you are going to support
that sort of scheme, I would suggest RAID1+0 (mirror first, then stripe)
rather than RAID0+1 (stripe first, then mirror). It's safer/slightly more
redundant, and FAR easier to extend on-the-fly (add two more drives,
reconfigure the RAID volume, extend the FS).

/pg
-- 
Peter Green : Gospel Communications Network, SysAdmin : [EMAIL PROTECTED]
---
If life deals you lemons, why not go kill someone with the lemons. (maybe by 
shoving them down his throat).
 (Jack Handey)