Re: NFS caching issues on 2.4

2001-06-06 Thread Trond Myklebust

> " " == Phil Oester <[EMAIL PROTECTED]> writes:

 > I've stumbled upon a wierd NFS caching issue on 2.4 which does
 > not seem to exist in 2.2.  Our www docroot is NFS mounted on a
 > NetApp 760.  We have a cron job which make changes to the
 > index.html every 5 minutes.

 > Recently, we upgraded all the web servers to 2.4 and noticed
 > that there were big delays in seeing these 5 minute updates.
 > Yet, an ls -l in the nfs directory on each of the servers
 > clearly shows the new timestamp.  However, a cat of the file
 > shows that it is the old version (sometimes up to 1 hour old).
 > I was using NFSv3, so decided to try NFSv2, but got the same
 > results.

 > I reverted to 2.2.19 on the boxes (which are RedHat 7.1
 > incidentally), and these problems went away.

 > Any ideas why this is happening?

The main suspect would be if you're mmapping the file while the cron
job updates your file on the server. This would means that the cache
invalidation breaks (see all the conditions in
invalidate_inode_pages()).

Cheers,
  Trond
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



NFS caching issues on 2.4

2001-06-06 Thread Phil Oester

I've stumbled upon a wierd NFS caching issue on 2.4 which does not seem to
exist in 2.2.  Our www docroot is NFS mounted on a NetApp 760.  We have a
cron job which make changes to the index.html every 5 minutes.

Recently, we upgraded all the web servers to 2.4 and noticed that there were
big delays in seeing these 5 minute updates.  Yet, an ls -l in the nfs
directory on each of the servers clearly shows the new timestamp.  However,
a cat of the file shows that it is the old version (sometimes up to 1 hour
old).  I was using NFSv3, so decided to try NFSv2, but got the same results.

I reverted to 2.2.19 on the boxes (which are RedHat 7.1 incidentally), and
these problems went away.

Any ideas why this is happening?

-Phil Oester

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



NFS caching issues on 2.4

2001-06-06 Thread Phil Oester

I've stumbled upon a wierd NFS caching issue on 2.4 which does not seem to
exist in 2.2.  Our www docroot is NFS mounted on a NetApp 760.  We have a
cron job which make changes to the index.html every 5 minutes.

Recently, we upgraded all the web servers to 2.4 and noticed that there were
big delays in seeing these 5 minute updates.  Yet, an ls -l in the nfs
directory on each of the servers clearly shows the new timestamp.  However,
a cat of the file shows that it is the old version (sometimes up to 1 hour
old).  I was using NFSv3, so decided to try NFSv2, but got the same results.

I reverted to 2.2.19 on the boxes (which are RedHat 7.1 incidentally), and
these problems went away.

Any ideas why this is happening?

-Phil Oester

-
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/



Re: NFS caching issues on 2.4

2001-06-06 Thread Trond Myklebust

   == Phil Oester [EMAIL PROTECTED] writes:

  I've stumbled upon a wierd NFS caching issue on 2.4 which does
  not seem to exist in 2.2.  Our www docroot is NFS mounted on a
  NetApp 760.  We have a cron job which make changes to the
  index.html every 5 minutes.

  Recently, we upgraded all the web servers to 2.4 and noticed
  that there were big delays in seeing these 5 minute updates.
  Yet, an ls -l in the nfs directory on each of the servers
  clearly shows the new timestamp.  However, a cat of the file
  shows that it is the old version (sometimes up to 1 hour old).
  I was using NFSv3, so decided to try NFSv2, but got the same
  results.

  I reverted to 2.2.19 on the boxes (which are RedHat 7.1
  incidentally), and these problems went away.

  Any ideas why this is happening?

The main suspect would be if you're mmapping the file while the cron
job updates your file on the server. This would means that the cache
invalidation breaks (see all the conditions in
invalidate_inode_pages()).

Cheers,
  Trond
-
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/