Hello everyone,
I was hoping to get recommendations on the proper way to test a RAID 1 hardware configuration. The controller is an Adaptec 2200S. I found an article, but not for this controller, that suggests to power off the system, pull one of the drives, boot the OS and power off again. Would this be the way to go?
Also any suggestions on what to use to monitor it as well? Can't get smartmontools to work.
Thank you and have a nice day.
Dianne
Hi,
On 05/27/2009 06:17 PM, Dianne Yumul wrote:
I was hoping to get recommendations on the proper way to test a RAID 1 hardware configuration. The controller is an Adaptec 2200S. I found an article, but not for this controller, that suggests to power off the system, pull one of the drives, boot the OS and power off again. Would this be the way to go?
That sounds quite silly - also going through reboots means downtime, isnt that the sort of thing that raid1 was designed to protect against anyway.
Also any suggestions on what to use to monitor it as well? Can't get smartmontools to work.
Have you looked into adaptec supplied management s/w ? In pretty much ever case with such hba's the most functional way to look at state and do any management on raid tends to be from vendor supplied s/w
Karanbir Singh schrieb:
Have you looked into adaptec supplied management s/w ? In pretty much ever case with such hba's the most functional way to look at state and do any management on raid tends to be from vendor supplied s/w
3Ware and Areca supply quite useful utilities that also work in Linux very well.
Adaptec - I don't know. They aquired a lot of companies over the years and thus the quality of the management-apps is/was sometimes questionable - and you never knew which one was good and which one wasn't as it varied between different revisions of the same hardware (which might have a similar name but a totally different tech inside...).
I've got lot's of Adaptec-SCSI gear at home - just no use for it anymore, because SCSI-drives don't make sense anymore in a desktop. It's not even the capacity - it's the noise.
Rainer
On 05/28/2009 02:11 PM, Rainer Duffner wrote:
3Ware and Areca supply quite useful utilities that also work in Linux very well.
yes, some of the LSI and HP kit is also quite handy to work with. I was going to setup a section on these HBA's in the wiki, would you be able to help on that ? If so, drop an email to the centos-docs list with your WikiName and we can take it from there. I'll do a template of what is in my mind in a few minutes time.
Thanks for the help Karanbir and Rainer.
That sounds quite silly - also going through reboots means downtime, isnt that the sort of thing that raid1 was designed to protect against anyway.
I guess it would be silly :). I just wanted to make sure it was doing what it ought too.
Have you looked into adaptec supplied management s/w ? In pretty much ever case with such hba's the most functional way to look at state and do any management on raid tends to be from vendor supplied s/w
Adaptec - I don't know. They aquired a lot of companies over the years and thus the quality of the management-apps is/was sometimes questionable - and you never knew which one was good and which one wasn't as it varied between different revisions of the same hardware (which might have a similar name but a totally different tech inside...).
I checked the Adaptec site again and found a link to the Adaptec Storage Manager. I don't know how I missed that, thanks for the push in the right direction. Hopefully it works.
Thanks,
dianne
on 5-28-2009 9:25 AM Dianne Yumul spake the following:
Thanks for the help Karanbir and Rainer.
That sounds quite silly - also going through reboots means downtime, isnt that the sort of thing that raid1 was designed to protect against anyway.
I guess it would be silly :). I just wanted to make sure it was doing what it ought too.
Have you looked into adaptec supplied management s/w ? In pretty much ever case with such hba's the most functional way to look at state and do any management on raid tends to be from vendor supplied s/w
Adaptec - I don't know. They aquired a lot of companies over the years and thus the quality of the management-apps is/was sometimes questionable - and you never knew which one was good and which one wasn't as it varied between different revisions of the same hardware (which might have a similar name but a totally different tech inside...).
I checked the Adaptec site again and found a link to the Adaptec Storage Manager. I don't know how I missed that, thanks for the push in the right direction. Hopefully it works.
It worked for me when I ran a 2800S, but it is far from a light weight app.
Dianne Yumul wrote:
I checked the Adaptec site again and found a link to the Adaptec Storage Manager. I don't know how I missed that, thanks for the push in the right direction. Hopefully it works.
Thanks,
dianne
Hi,
I use the Adaptec Storage manager under CentOS 5 (with Adaptec 3405 HBA). It seems to be decent (functionality wise) but i saw that it works under Java... My deception was great when i saw that. I had numerous problems with Java and management tools on the past. I just hope i wont't have to fight with Java all over again...
Guy Boisvert, ing. IngTegration inc.
Thanks Scott and Guy, I appreciate the advice/warning.
I can't download it at the moment because the Adaptec site requires the serial number. I have to reboot and get it from the SMOR (Storage Manager on ROM) utility. Then I'll have to see if it's worth the trouble to install and run :)
Would you have any suggestions on how I would test the RAID 1 configuration? (Sorry if you already saw the question on the previous posts)
Thanks,
dianne
Dianne Yumul wrote: ...
Would you have any suggestions on how I would test the RAID 1 configuration? (Sorry if you already saw the question on the previous posts)
Dianne,
If you feel you must test the functionality of the ability of the RAID1 to recover from a failed drive, the power down, remove drive, boot, power down, replace drive, boot process will test the ability of the system to re-sync the drives. A more rigorous test would be to zero out the partition table of the removed drive, or to use a new blank drive to test the recovery; however, what I think Karanbir meant by "silly" is that all these tests simply confirm that the RAID is working as designed. If it doesn't, then you should have bought different hardware to start with.
Phil
On May 28, 2009, at 1:00 PM, Phil Schaffner wrote:
If you feel you must test the functionality of the ability of the RAID1 to recover from a failed drive, the power down, remove drive, boot, power down, replace drive, boot process will test the ability of the system to re-sync the drives. A more rigorous test would be to zero out the partition table of the removed drive, or to use a new blank drive to test the recovery; however, what I think Karanbir meant by "silly" is that all these tests simply confirm that the RAID is working as designed. If it doesn't, then you should have bought different hardware to start with.
Thanks Phil for the clarification, felt like the light bulb suddenly turned on in my head. I should be more concerned at monitoring the health of the hard drives then.
dianne
Phil Schaffner wrote:
Dianne Yumul wrote: ...
Would you have any suggestions on how I would test the RAID 1 configuration? (Sorry if you already saw the question on the previous posts)
Dianne,
If you feel you must test the functionality of the ability of the RAID1 to recover from a failed drive, the power down, remove drive, boot, power down, replace drive, boot process will test the ability of the system to re-sync the drives. A more rigorous test would be to zero out the partition table of the removed drive, or to use a new blank drive to test the recovery; however, what I think Karanbir meant by "silly" is that all these tests simply confirm that the RAID is working as designed. If it doesn't, then you should have bought different hardware to start with.
The thing you need to know about RAIDs at runtime is whether or not one or more of the drives have already failed since their job is to hide this fact from you but you still need to replace it before you lose the other one and your data. Most systems have an error light or something that doesn't help much in a remote data center - so you need some hardware-specific monitor tool to report the status.
On May 28, 2009, at 2:01 PM, Les Mikesell wrote:
The thing you need to know about RAIDs at runtime is whether or not one or more of the drives have already failed since their job is to hide this fact from you but you still need to replace it before you lose the other one and your data...
Got it. Thank you Les, I appreciate the explanation.
dianne
On 05/28/2009 07:43 PM, Dianne Yumul wrote:
I can't download it at the moment because the Adaptec site requires the serial number. I have to reboot and get it from the SMOR (Storage Manager on ROM) utility. Then I'll have to see if it's worth the trouble to install and run :)
what sort of license is this distributed under ?
On May 28, 2009, at 1:22 PM, Karanbir Singh wrote:
what sort of license is this distributed under ?
I'm sorry but I'm not sure. It's software that supposedly comes with it when you buy it. But the server was assembled by somebody else and they probably neglected to include the CD when they shipped the server to us. Here's the page with the download link.
http://www.adaptec.com/en-US/speed/raid/aac/sm/asm-linux_v2_12(922) _rpm.htm
I'm going to give it a try but don't know when because I would have to bring the server down just to get the serial number.
Thanks Karanbir and everyone for all the help.
dianne
on 5-28-2009 1:56 PM Dianne Yumul spake the following:
On May 28, 2009, at 1:22 PM, Karanbir Singh wrote:
what sort of license is this distributed under ?
I'm sorry but I'm not sure. It's software that supposedly comes with it when you buy it. But the server was assembled by somebody else and they probably neglected to include the CD when they shipped the server to us. Here's the page with the download link.
http://www.adaptec.com/en-US/speed/raid/aac/sm/asm-linux_v2_12(922)_rpm.htm
I'm going to give it a try but don't know when because I would have to bring the server down just to get the serial number.
Thanks Karanbir and everyone for all the help.
dianne
Sent you a message offlist on this...
On Thu, 2009-05-28 at 09:25 -0700, Dianne Yumul wrote:
Thanks for the help Karanbir and Rainer.
That sounds quite silly - also going through reboots means downtime, isnt that the sort of thing that raid1 was designed to protect against anyway.
I guess it would be silly :). I just wanted to make sure it was doing what it ought too.
Have you looked into adaptec supplied management s/w ? In pretty much ever case with such hba's the most functional way to look at state and do any management on raid tends to be from vendor supplied s/w
Adaptec - I don't know. They aquired a lot of companies over the years and thus the quality of the management-apps is/was sometimes questionable - and you never knew which one was good and which one wasn't as it varied between different revisions of the same hardware (which might have a similar name but a totally different tech inside...).
I checked the Adaptec site again and found a link to the Adaptec Storage Manager.
Be aware the Storage Manager does not support CentOS 5.
I don't know how I missed that, thanks for the push in the right direction. Hopefully it works.
Thanks,
dianne _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
-
Dianne Yumul -
Guy Boisvert -
JohnS -
Karanbir Singh -
Les Mikesell -
Phil Schaffner -
Rainer Duffner -
Scott Silva