[prev in list] [next in list] [prev in thread] [next in thread] 

List:       evms-devel
Subject:    Re: [Evms-devel] RAID1 recovery?
From:       "Steve Pratt" <slpratt () us ! ibm ! com>
Date:       2002-05-21 23:39:03
[Download RAW message or body]


Stefan Gybas wrote:
>Hi Steve,

>thanks for your quick response and help.

No problem.

>>>I then tried to reactivate hdc2 using evmsgui. I've found an option in
>>>"Modify Properties.. " for the md0 region where hdc2 was listed as spare
>>>object: "Activate selected Spare Object". But this did not work like I
>>>exepcted. /proc/evms/mdstat now shows this for md0:
>>
>> This was the correct thing to do.

>How could I do this with evms from the command line in case this happens
>in real life? I guess I have to use a set command, but which one?

That would be a good guess.  I had to look it up myself :-)  Unfortunately
it appears that the ListOptions query in the command line is broken and
does not return the options valid for a set command.

The command would be something like:

set:md/md0,availdisks=sdc4,addspare=TRUE

I'll get with Ben in the morning to see what is wrong with the query.


>> If you do a "display details" on md0 what do you see? It should list all
of
>> the disks(object) that make up md0 as well as the current state of the
>> devices.  Use the "more..." buttons to see this in the gui.

Here's all the information from evmsgui:

> info deleted......

Yes, the array has a problem.

I was able to reproduce the error here.  It stems from using the utime
field in the Engine instead of the event counters for freshenss checking.
This would have been OK, except there seems to be a problem where sometimes
the kernel uses Local Time and Sometimes GMT.  This caused the timestamps
to be out of sync and we picked the failed disk as the freshest superblock.
Things kind of went downhill from there.  I have checked in a new
md_discover.c to CVS.  We should be releasing a 1.1 patch in a week or 2 if
you want to wait, or if you want I can create you a patch file tomorrow.


>I also got the "deleteing duplicate reference to ..." for some other
>regions. Does this means there's something wrong with my setup? The
>complete disk was set up with EVMS only, i.e. no compatibility MD, LVM
>or volumes.

No, this is normal, but probably confusing.  I'll get with Mark to see
about cleaning this up.

>[I'm posting the logs to the list so people can search for keywords in
>the archive if they have the same problem. I hope this is ok.]

That is fine, as long as they don't get too big.  (These were fine)

Steve




_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm

_______________________________________________
Evms-devel mailing list
Evms-devel@lists.sourceforge.net
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel
[prev in list] [next in list] [prev in thread] [next in thread]