Am Freitag 18 Januar 2008 17:49:15 schrieb Anand Avati:
Sascha,
the reason why 1.3.0pre4 might be faster would not be because of the
missing namespace, but most likely because of missing self-heal. can
you try with 'option self-heal off' in the unify section?
may ask again, any idea why the
Hi Sascha
I have a few ideas you can try to figure out a bit more wahts happening
Maybe there is locking on the apache as it traverses files.
Asking the devels:
Is locking implemented on files (posix-locks over posix-storage) or does unify
locks also files on namespace?
Are there any
On Sat, 19 Jan 2008 18:03:45 +0530, Angel [EMAIL PROTECTED] wrote:
Hi all
Whats IDE are you using for development of GlusterFS? Kdevelop, Eclipse,
Turbo C++2.0 :-P
All of use the one and only Emacs :)
PS: Are you going to expand documentation about internals, on the next
release?? Hackers
Hi,
don't know if this is already fixed, just in case here are two things I
observered with patch-628:
~# LD_PRELOAD=/usr/lib/glusterfs/1.3.8/xlator/performance/booster.so
nginx
nginx: symbol lookup
error: /usr/lib/glusterfs/1.3.8/xlator/performance/booster.so:
undefined symbol:
Hi
I found on posix writev that stbuf is local , but a ptr to it is returned to
caller.
int32_t
posix_writev (call_frame_t *frame,
xlator_t *this,
fd_t *fd,
struct iovec *vector,
int32_t count,
off_t offset)
{
int32_t
Hi August
One direct source to learn what things happen is looking at the same function
on the posix-storage module.
that's how posix-storage process create
...
SET_FS_UID (frame-root-uid, frame-root-gid);
if (!flags) {
_fd = open (real_path,
O_CREAT|O_RDWR|O_EXCL,
Hi glusterfs hackers,
I am trying to write a simple xlator similar to the fixed-id one that comes
with the source code.
The fixed-id xlator passes all calls through to the underlying volume and
then mangles the uid and gid of the stat structures on the way back to the
client, so what you end up
Hi, all
I just send you my QUOTA xlator, it is still VERY ALPHA code, but yet usefull
to allow devs and people
comment and report on funcionality, errors, coding guidelines and so.
Quota is a dirty hack over trace and need a lot cleaning :-) but patch compiles
cleanly againt tla636
It offers
Am Samstag 19 Januar 2008 21:46:26 schrieb Sascha Ottolski:
Am Samstag 19 Januar 2008 14:16:31 schrieb Sascha Ottolski:
For me, I'm only serving small static image files that live on a
gluster mount. And find it strange and unexpected that apache1
outperforms apache2, nginx and lighttpd so
now, another astonishing observation: if I enable the io-cache, it has a
good effect for apache1 (almost doubles the requests/second), but
almost none for apache2, nginx and lighttpd.
could this help to understand more about the performance differences?
io-cache should have helped all of
Am Sonntag 20 Januar 2008 00:59:12 schrieb Anand Avati:
now, another astonishing observation: if I enable the io-cache, it
has a good effect for apache1 (almost doubles the requests/second),
but almost none for apache2, nginx and lighttpd.
could this help to understand more about the
Am Samstag 19 Januar 2008 11:03:43 schrieb Sascha Ottolski:
Am Freitag 18 Januar 2008 17:49:15 schrieb Anand Avati:
Sascha,
the reason why 1.3.0pre4 might be faster would not be because of
the missing namespace, but most likely because of missing
self-heal. can you try with 'option
Sascha,
Please run by removing AFR (make AFR's first subvolume a direct subvolume
of unify). AFR can potentially flush io-cache pages while doing read
scheduling across subvolumes (by returning inconsistent mtimes with the read
buffer). we are working on fixing this.
avati
2008/1/20, Sascha
Am Sonntag 20 Januar 2008 01:14:06 schrieb Anand Avati:
can you please share your spec files with glusterfs.pastebin.com ?
my pleasure: http://glusterfs.pastebin.com/m8820300
Thanks for your time,
Sascha
thanks,
avati
2008/1/20, Sascha Ottolski [EMAIL PROTECTED]:
Am Samstag 19 Januar
14 matches
Mail list logo