"It" isn't a drive.  As I mentioned, it is a RAID array managed by
mdadm.  There are ten 1T drives in the RAID 6 array.  A couple of t he
drivesa re Seagate drives, four are Western Digital drives, and the
rest are Hitachi.

On Jan 18, 7:39 am, "Daniel Eggleston" <[email protected]> wrote:
> What kind of hard drive is it? Seagate is having some firmware struggles
> lately...
>
>
>
> On Sun, Jan 18, 2009 at 8:34 AM, lrhorer <[email protected]> wrote:
>
> > I am having a problem with performance of the drive system on a Debian
> > "Lenny" distro.  This is running RAID under mdadm with a reiserfs file
> > system on a raw partition.  The kernel is 2.6.26-1-amd64 running on an
> > AMD Athlon 64 x 2 CPU with 4G of RAM.  Ordinarily, performance is
> > quite good, and I have transferred files as fast as 480 Mbps upon
> > occasion.  Frequently, however, drive access will drop to dead zero
> > for about 40 seconds.  The trigger event seems to be file creation,
> > and I don't believe it usually happens unless a file is being
> > created.  It does not happen every time a file is created, but for
> > example I have a script which performs a number of tests and checks
> > and at the end creates a log file.  When the script reaches that
> > point, it often hangs, and if there are any network file transfers
> > going on in the background (there usually are), the network
> > utilization drops right off to almost zilch.  Another common trigger
> > is editing a video on one of the served Windows workstations.  At some
> > point the edit parameters are saved in a cut file.  When the user
> > first selects the "Save Project" utility, the file write often hangs
> > for about 40 seconds.  Sometimes Windows will report the drive is
> > full, or a timeout has occurred.  I do not know for a fact that drive
> > - drive transfers are suspended, but  NFS, SAMBA, FTP, and other
> > network transfers in both directions are abrupted for a period of
> > about 40 seconds.  Non-drive related transfers, or transfers to
> > devices other than the RAID array are not impacted.  There is nothing
> > in any of the logs to suggest anything at all has happened.
> > Interestingly enough, although Windows sometimes complains, ordinary
> > TCP transfers do not time out, despite the fact not a single byte
> > transfers in either direction during the event.
>
> > What can I do to try to find out what is hanging and why?
>
> --
>
>           Daniel
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Linux Users Group.
To post a message, send email to [email protected]
To unsubscribe, send email to [email protected]
For more options, visit our group at 
http://groups.google.com/group/linuxusersgroup
-~----------~----~----~----~------~----~------~--~---

Reply via email to