SCIENTIFIC-LINUX-USERS Archives

July 2014

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Andras Horvath <[log in to unmask]>
Reply To:
Andras Horvath <[log in to unmask]>
Date:
Wed, 2 Jul 2014 09:14:47 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (30 lines)
On Tue, 1 Jul 2014 14:47:27 -0700
"Patrick J. LoPresti" <[log in to unmask]> wrote:

> On Tue, Jul 1, 2014 at 1:28 PM, Andras Horvath <[log in to unmask]> wrote:
> > On Tue, 1 Jul 2014 13:19:10 -0700
> >> Something like this is more sane:
> >>
> >>   sysctl -w vm.dirty_background_bytes=67108864
> >>   sysctl -w vm.dirty_background_bytes=134217728
> 
> Cut&paste error there. That second line should read
> "vm.dirty_bytes=134217728". The idea is to block the writer if the
> page cache gets sufficiently backed up.
> 
> You can also try turning down vm.dirty_writeback_centisecs and
> dirty_expire_centisecs, within reason... That should also help to keep
> the writes flowing.
> 
> Sorry this didn't help. Good luck.
> 
>  - Pat

Anyway I tried to raise those values high again and it seemed to have enter that hanging state at once. Interesting.

I also checked the SATA possibility and it's working just fine if no USB is involved in the story. Since I have no more idea, I'll stick with the SATA version.

Thanks.

Andras

ATOM RSS1 RSS2