On Tue, Feb 12, 2008 at 02:39:28PM +0100, markus neis wrote:
After troubleshooting my system I found out, that there were also troubles
with one of my NIC's. This problem is fixed and posix locking = no is
working pretty good with gfs now! I hadn't the possibility to test this
config with
After troubleshooting my system I found out, that there were also troubles
with one of my NIC's. This problem is fixed and posix locking = no is
working pretty good with gfs now! I hadn't the possibility to test this
config with windows clients and I think I won't have it anymore. Are there
any
On Sun, Feb 10, 2008 at 01:20:20AM +0100, Markus Neis wrote:
here's my smb.conf if you would be so nice ... ;-)
No. Analyzing performance problems can't be done by looking
at config files only. You have to sit at the console of that
box and see the dynamic behaviour :-)
Volker
That's true, but I still wonder why there are no performance problems when I
set locking = no !
_
Unbegrenzter Speicherplatz für Ihr E-Mail Postfach? Jetzt aktivieren!
http://www.digitaledienste.web.de/freemail/club/lp/?lp=7
--
Volker Lendecke wrote:
On Sun, Feb 10, 2008 at 12:51:24AM +0100, Markus Neis wrote:
Ok I understand. People that earn more money than me made
this decision ;-) But I wonder why Redhat offers the
possibility in their cluster suite to setup samba services
on top of gfs. This should be a
On Sun, Feb 10, 2008 at 10:56:30AM -0500, Ross S. W. Walker wrote:
You can perfectly fine use samba on top of gfs, as long as
you only share your data from a single node or (more
precisely) make sure that every directory is only shared via
a single node. Different directories can be shared
Hi there,I run samba as a PDC and tried to make this PDC high available with
redhat cluster suite and gfs. I experienced the following problem while
doing this:
If I set the option locking = no in smb.conf it takes about 4 minutes to
copy a file of 1GB size. If I set locking = yes it takes about 1
Hi there,
I run samba as a PDC and tried to make this PDC high available with redhat
cluster suite and gfs. I experienced the following problem while doing this:
If I set the option locking = no in smb.conf it takes about 4 minutes to copy a
file of 1GB size. If I set locking = yes it takes
On Feb 9, 2008 8:49 AM, markus neis [EMAIL PROTECTED] wrote:
Hi there,I run samba as a PDC and tried to make this PDC high available
with
redhat cluster suite and gfs. I experienced the following problem while
doing this:
If I set the option locking = no in smb.conf it takes about 4 minutes
On Feb 9, 2008 3:01 PM, markus neis [EMAIL PROTECTED] wrote:
i set oplocks = yes , kernel oplocks = yes and as I said locking = yes,
but this slows down everything
OK, from what I gather (which very well could be inaccurate), it looks like
you might be stuck on a spinlock timeout on a
i set oplocks = yes , kernel oplocks = yes and as I said locking = yes, but
this slows down everything
_
Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
Damn! this doesn't sound good. I hope somebody else can refute what you say ;-)
gfs shouldn't be that slow. I'm really confused.
___
GRATIS: Movie-FLAT. Jetzt freischalten!
http://freemail.web.de/club/maxdome.htm
--
To unsubscribe from this list go to the
On Sat, Feb 09, 2008 at 11:31:59PM +0100, Markus Neis wrote:
Damn! this doesn't sound good. I hope somebody else can
refute what you say ;-)
gfs shouldn't be that slow. I'm really confused.
No offense intended, but Scott's description is not really
correct. The only parameter that should
Volker Lendecke wrote:
On Sat, Feb 09, 2008 at 11:31:59PM +0100, Markus Neis wrote:
Damn! this doesn't sound good. I hope somebody else can
refute what you say ;-)
gfs shouldn't be that slow. I'm really confused.
No offense intended, but Scott's description is not really
correct.
ok, but i heard people say that ctdb isn't for productional use. can you
confirm that volker?
Is it also a good idea to save the locking information on the gfs filesystem?
Bis 50 MB Dateianhänge? Kein Problem!
On Sun, Feb 10, 2008 at 12:17:02AM +0100, Markus Neis wrote:
ok, but i heard people say that ctdb isn't for
productional use. can you confirm that volker?
Well, it *is* used in production. It's not trivial to set
up, but it works.
Is it also a good idea to save the locking information on
the
Ok I understand. People that earn more money than me made this decision ;-) But
I wonder why Redhat offers the possibility in their cluster suite to setup
samba services on top of gfs. This should be a known problem then. However, as
I understand you its not a good idea to use samba with gfs
On Sat, Feb 09, 2008 at 06:07:36PM -0500, Scott Lovenberg wrote:
Just to clarify, the locking semantics (regardless of type) do not
propagate down to the kernel smb module, but rather pass to the
underlying file system (which in turn propagates to its own kernel
module)? Thanks, Volker.
here's my smb.conf if you would be so nice ... ;-)
[global]
dos charset = 850
unix charset = CP850
display charset = LOCALE
workgroup = FILESRV
realm =
netbios name = FILESRV
netbios aliases = austauschsrv, fachaustsrv, grplwsrv
On Sun, Feb 10, 2008 at 01:00:07AM +0100, Markus Neis wrote:
I put the tdb files out of gfs only some data files are
now on gfs, but it's the same behavior. ;-(
If it's still slow with posix locking = no, and tdb files
out of GFS, then more detailed analysis is necessary. This
will be tough
On Sun, Feb 10, 2008 at 12:51:24AM +0100, Markus Neis wrote:
Ok I understand. People that earn more money than me made
this decision ;-) But I wonder why Redhat offers the
possibility in their cluster suite to setup samba services
on top of gfs. This should be a known problem then.
However,
I put the tdb files out of gfs only some data files are now on gfs, but it's
the same behavior. ;-(
Bis 50 MB Dateianhänge? Kein Problem!
http://www.digitaledienste.web.de/freemail/club/lp/?lp=7
--
To unsubscribe from this list go to the
On Sun, Feb 10, 2008 at 01:08:04AM +0100, Volker Lendecke wrote:
You can perfectly fine use samba on top of gfs, as long as
you only share your data from a single node or (more
precisely) make sure that every directory is only shared via
a single node. Different directories can be shared via
23 matches
Mail list logo