[Beowulf] Big storage

Bruce Allen ballen at gravity.phys.uwm.edu
Thu Aug 30 02:37:16 PDT 2007


Hi Andrew,

I've corresponded with Neil Brown about this point in the past.  He is 
aware of this shortcoming.  If you look at his 'to-do' list (here is the 
URL: http://neil.brown.name/blog/20050727141521) you will find the item: 
"Don't kick drives on read errors".  Just after that is "background 
check/repair" where it says "It would be good to also do background read 
checks."

In the next year I might look for a good student in Hannover who wants to 
do this, and help them modify the code.  Any students from U. of Hannover 
reading this who are interested, please drop by my office for a chat!

Cheers,
 	Bruce

On Thu, 30 Aug 2007, Andrew Walrond wrote:

> Bruce Allen wrote:
>>
>> Yes, precisely.
>>
>> When I buy RAID controllers, I put this requirement directly in my bid
>> specifications.  I say something like the following:
>>
>> ------------------------------------------------------------------
>>
>> During a READ operation, if the RAID controller finds an unreadable
>> (uncorrectable) disk sector, then it will immediately reconstruct the
>> missing data for that sector using redundant data from the rest of the
>> array, and WRITE that data to the unreadable (uncorrectable) sector to
>> force sector reallocation if needed by the disk.
>>
>> In addition, the RAID controller will perform a continous or regular (at
>> least daily) background scan of all disk sectors to identify and repair
>> any unreadable sectors as described above.
>>
>> ------------------------------------------------------------------
>>
>> I think it is a mistake to purchase a RAID controller without these
>> features.  The absence of these features is the main reason that I don't
>> think Linux software RAID is very good.
>>
>
> Ouch! I'd mentally filed linux software raid and good and trustworthy
> quite a while ago and use it heavily. Presumably the maintainers (Neil
> Brown? or is he just the mdadm maintainer) are aware of this lack/criticism?
>
> As presented, it certainly seems like features that ought to be added.
> It would be interesting to cc LKML for comment.
>
> Andrew Walrond
>



More information about the Beowulf mailing list