[Ocfs-users] Re: Clarification on 1.0.14-1

[Sunil Mushran]

Oh yes. The reason cp is slow on ocfs because it not only uses the
default 512byte blocksize, it also floods the buffercache. This is
especially true when dealing with large files.

With --o_direct in the modified fileutils/coreutils, one can not only 
specify
the blocksize, but also not flood the buffercache. win-win.

<<<<<<<<<<
There is not a lot of info on the modified coreutils out there, so any
thoughts are again appreciated.
>>>>>>>>>>>

We have updated the man pages, but guess could do with adding
some info on the web page itself. I'll look into that.

Bruce Holzrichter wrote:

>>Have you tried archiving to another ocfs volume?
>>
>>The reason I bring this up is because you can avoid flooding
>>the buffer cache by using ocfs (o_direct) instead of ext2 (buffered).
>>    
>>
>
>Thanks again for your replies.
>
>We do a lot of regular file moves, backup to tape, etc, to keep more
>than a days archivelog backups on those partitions.  
>
>There was considerable slowness with regular file utility mv,cp, etc,
>would using the Oracle coreutils with the o_direct make a difference
>here?
>
>There is not a lot of info on the modified coreutils out there, so any
>thoughts are again appreciated.
>
>Bruce H.
>
>
>
>
>
>
>  
>