Help with failed RAID1 on XW6400

[sycomp]

I have an XW6400 Workstation configured with 2 RAID1 Enterprise drives on an LSI SAS-3041E-HP card. For approximately 3 days, the LSI card has been warning me that the drive at Port 0 has been getting ready to fail. The array is in a degraded state.

The machine is a business machine that I need for daily business operations, so while I'm waiting on the replacement drive to arrive, I've been using it. Today, it was running slow and intermittently freezing, hence I figured the drive was the issue. I thought I'd just go ahead and unplug the drive and use the secondary drive since it's RAID 1.

No joy. The OS will no longer boot.

I'm a little confused...I thought the drives would be mirrored, hence, in the case of failure I could still use one drive until the other was replaced.

Where am I going wrong?

[DGroves]

lets start with the basics, are you sure it's a raid 1 array? check the LSI cards bios for the current array config do not accept any changes if offered

how did you remove the failed drive from the array? just removing it is a no no, and will cause problems.

please reinsert the drive

did you remove the failed drive or the working one? how did you determine which drive had failed?

the failed drive needs to be marked as such in the LSI cards bios before removal. and the replacemnet added to the current array via the LSI card bios 

have you read the LSI cards manual on how to replace a failed drive? if not do so now.

Hint,.......the how to  part should be read and data recovery procedures tested before you place data on a array i recomend you follow this policy in the future

https://usermanual.wiki/Lsi/LsiSas3041EUsersManual576179.697282750/help

[sycomp]

[EDIT: I appreciate the manual link you sent, but it doesn't seem to deal with replacing failed drives. Did you mean to link another manual by chance?]

Thanks for your reply.

lets start with the basics, are you sure it's a raid 1 array? check the LSI cards bios for the current array config do not accept any changes if offered

Yes certain.

how did you remove the failed drive from the array? just removing it is a no no, and will cause problems.

please reinsert the drive

Unplugged the cable.

did you remove the failed drive or the working one? how did you determine which drive had failed?

I removed the failed one. Using the LSI configuration utility, I identifed the port number attributed to the drive by the card. The ports are indentified on the card itself.

the failed drive needs to be marked as such in the LSI cards bios before removal. and the replacemnet added to the current array via the LSI card bios 

It looks like it already is...the bios configuration utility has labeled it as "failed," even though the machine was bootable. This is why I assumed the mirrored array would allow me to boot with only one drive.

Thanks for the link to manual!

Any idea why a known good mirrored drive won't boot?

[DGroves]

no, the LSI bios is saying  THE CURRENT DRIVE IN PORT X HAS FAILED,

[DGroves]

no, the LSI bios is saying  THE CURRENT DRIVE IN PORT X HAS FAILED,  THE FAILED DRIVE IS STILL PART OF THE ARRAY and is now missing from the array since you removed the drive

you must reinstall the failed drive, then enter the LSI bios and remove the failed drive from the current array

in a raid 1, this means you either break the array (IE-remove the raid 1 array) back to normal drives or  in a raid 5 you mark the failed drive as removed from the array before add in the replacement drive to the array

here's a better manual link

https://www.manualslib.com/manual/96619/Lsi-Lsisas3041e.html

[sycomp]

you must reinstall the failed drive, then enter the LSI bios and remove the failed drive from the current array

Gotcha...at that point the 2ndary drive should be bootable then, or am I down until the replacement arrives?

here's a better manual link

Actually that appears to be the same 46 page manual you linked previously. Thank you anyway!

If you have some extra time...what about when I get my replacement drive? I notice a "synchronize" option within the card's bios. Will there be any issues (e.g., wiping the drives) if I sync the new drive within the CARD's bios, rather than from Windows? Or would it be better to configure the array from the Windows utility, and is that even possible? Originally, I configured the array from the card's bios.

Many thanks.

[DGroves · Accepted solution]

yes, removing the failed drive from the array should allow the remaining drive to boot

i'm not sure how the HP firmware on this LSI card is setup, it will either remove the drive from the raid 1 leaving a degraded bootable array till you add in the replacement drive or if it simply breaks the raid1 back into std drives

the manual should cover this in detail

[sycomp]

the manual should cover this in detail

Hmm...I was hoping you'd have an idea...the manual doesn't appear to cover it at all.

I looked around here as well: https://support.hp.com/us-en/product/hp-xw6400-workstation/1844973/manuals

No joy for the required info there either.

Thanks for your help.

[sycomp]

Interesting...once I plugged the "bad" drive back in and rebooted, the card's bios recognized the drive and the status changed to "synchronizing."

I left it there for a bit and then decided to boot into Windows as a test. According to the MegaRAID utility, it's rebuilding the array now.

Just to be sure I'd like to test the drive for failure with Data Lifeguard since these are Western Digital drives. According to this document I'd need to remove the drive from the array first: https://support.wdc.com/knowledgebase/answer.aspx?ID=996

Just out of curiosity, is anyone familiar with a utility that can recognize/test the separate drives while they're in the array? I'd hate to break it again if it isn't necessary.

[DGroves]

are these drives Raid or NAS certified? consumer drives can sometime go offline if they encounter a "soft" error and pause to recover from it. during the pause the drive ignores any new commands till it finishes. the scsi controller asumes the drive has failed if it does not respond to it's query in the time limit alloted

soft errors are things like encountering a read or write issue and having to retry to see if sucessful and if not remapping the failed sector this takes more time than the controller allows. raid certified drives have firmware that limits the time a HD can spend on housekeeping which prevents this type of raid error