It looks like I was posting on the wrong mailing list. I thought
this mailing list includes developers. The experiment I did is
not for commercial purpose. The purpose of comparison is to
find the optimization opportunity of the entire software stack
on both linux and solaris.
As for this zfs
One of the glories of Solaris is that it is so very observable. Then
there are the many excellent blog posts, wiki entries, and books - some
or which are authored by contributors to this very thread - explaining
how Solaris works. But these virtues are also a snare to some, and it is
not
hi
you did not answer the question, what is the RAM of the server? how many
socket and core etc
what is the block size of zfs?
what is the cache ram of your san array?
what is the block size/strip size of your raid in san array? raid 5 or
what?
what is your test program and how (from what
On Mon, Mar 26, 2012 at 4:33 PM, casper@oracle.com wrote:
I'm migrating a webserver(apache+php) from RHEL to solaris. During the
stress testing comparison, I found under the same session number of client
request, CPU% is ~70% on RHEL while CPU% is full on solaris.
Which version of Solaris
2012-03-26 14:27, Aubrey Li wrote:
The php temporary folder is set to /tmp, which is tmpfs.
By the way, how much RAM does the box have available?
tmpfs in Solaris is backed by virtual memory.
It is like a RAM disk, although maybe slower than ramdisk
FS as seen in livecd, as long as there is
You care about #2 and #3 because you are fixated on a ZFS root
lock contention problem, and not open to a broader discussion
about what your real problem actually is. I am not saying there is
not lock contention, and I am not saying there is - I'll look at the
data later carefully later when I
On Mon, Mar 26, 2012 at 7:28 PM, Jim Klimov jimkli...@cos.ru wrote:
2012-03-26 14:27, Aubrey Li wrote:
The php temporary folder is set to /tmp, which is tmpfs.
By the way, how much RAM does the box have available?
tmpfs in Solaris is backed by virtual memory.
It is like a RAM disk,
On Mon, Mar 26, 2012 at 8:24 PM, Jim Mauro james.ma...@oracle.com wrote:
You care about #2 and #3 because you are fixated on a ZFS root
lock contention problem, and not open to a broader discussion
about what your real problem actually is. I am not saying there is
not lock contention, and I
I see nothing unusual in the lockstat data. I think you're barking up
the wrong tree.
-- richard
On Mar 25, 2012, at 10:51 PM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 1:19 PM, Richard Elling
richard.ell...@richardelling.com wrote:
Apologies to the ZFSers, this thread really belongs
As a random guess, try pointing PHP tmp directory to
/var/tmp (backed by zfs) and see if any behaviors change?
Good luck,
//Jim
Thanks for your suggestions. Actually the default PHP tmp directory
was /var/tmp, and I changed /var/tmp to /tmp. This reduced zfs
root lock contention
On Tue, Mar 27, 2012 at 1:15 AM, Jim Klimov j...@cos.ru wrote:
Well, as a further attempt down this road, is it possible for you to rule
out
ZFS from swapping - i.e. if RAM amounts permit, disable the swap at all
(swap -d /dev/zvol/dsk/rpool/swap) or relocate it to dedicated slices of
same or
On Mar 24, 2012, at 10:29 PM, Aubrey Li wrote:
Hi,
I'm migrating a webserver(apache+php) from RHEL to solaris. During the
stress testing comparison, I found under the same session number of client
request, CPU% is ~70% on RHEL while CPU% is full on solaris.
After some investigation, zfs
On Sun, Mar 25, 2012 at 3:55 PM, Richard Elling
richard.ell...@richardelling.com wrote:
On Mar 24, 2012, at 10:29 PM, Aubrey Li wrote:
Hi,
I'm migrating a webserver(apache+php) from RHEL to solaris. During the
stress testing comparison, I found under the same session number of client
This is the wrong forum for general purpose performance tuning. So I won't
continue this much farther. Notice the huge number of icsw, that is a bigger
symptom than locks.
-- richard
On Mar 25, 2012, at 6:24 AM, Aubrey Li wrote:
SET minf mjf xcal intr ithr csw icsw migr smtx srw syscl
On Mon, Mar 26, 2012 at 12:48 AM, Richard Elling
richard.ell...@richardelling.com wrote:
This is the wrong forum for general purpose performance tuning. So I won't
continue this much farther. Notice the huge number of icsw, that is a
bigger
symptom than locks.
-- richard
thanks anyway,
On 3/25/12 10:25 AM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 12:48 AM, Richard Elling
richard.ell...@richardelling.com wrote:
This is the wrong forum for general purpose performance tuning. So I won't
continue this much farther. Notice the huge number of icsw, that is a
bigger
symptom than
On Mar 25, 2012, at 10:25 AM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 12:48 AM, Richard Elling
richard.ell...@richardelling.com wrote:
This is the wrong forum for general purpose performance tuning. So I won't
continue this much farther. Notice the huge number of icsw, that is a
bigger
On Mon, Mar 26, 2012 at 2:10 AM, zfs user zf...@itsbeen.sent.com wrote:
On 3/25/12 10:25 AM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 12:48 AM, Richard Elling
richard.ell...@richardelling.com wrote:
This is the wrong forum for general purpose performance tuning. So I
won't
continue this
On Mon, Mar 26, 2012 at 2:58 AM, Richard Elling
richard.ell...@richardelling.com wrote:
On Mar 25, 2012, at 10:25 AM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 12:48 AM, Richard Elling
richard.ell...@richardelling.com wrote:
This is the wrong forum for general purpose performance tuning. So I
On Mon, Mar 26, 2012 at 2:13 AM, Aubrey Li aubrey...@gmail.com wrote:
The problem is, every zfs vnode access need the **same zfs root**
lock. When the number of
httpd processes and the corresponding kernel threads becomes large,
this root lock contention
becomes horrible. This situation does
If you're chasing CPU utilization, specifically %sys (time in the kernel),
I would start with a time-based kernel profile.
#dtrace -n 'profile-997hz /arg0/ { @[stack()] = count(); } tick-60sec {
trunc(@, 20); printa(@0; }'
I would be curious to see where the CPU cycles are being consumed first,
On Mon, Mar 26, 2012 at 3:22 AM, Fajar A. Nugraha w...@fajar.net wrote:
I have ever not seen any issues until I did a comparison with Linux.
So basically you're comparing linux + ext3/4 performance with solaris
+ zfs, on the same hardware? That's not really fair, is it?
If your load is
On Mon, Mar 26, 2012 at 4:18 AM, Jim Mauro james.ma...@oracle.com wrote:
If you're chasing CPU utilization, specifically %sys (time in the kernel),
I would start with a time-based kernel profile.
#dtrace -n 'profile-997hz /arg0/ { @[stack()] = count(); } tick-60sec {
trunc(@, 20); printa(@0;
On Mar 25, 2012, at 6:51 PM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 4:18 AM, Jim Mauro james.ma...@oracle.com wrote:
If you're chasing CPU utilization, specifically %sys (time in the kernel),
I would start with a time-based kernel profile.
#dtrace -n 'profile-997hz /arg0/ { @[stack()] =
On Mon, Mar 26, 2012 at 11:34 AM, Richard Elling
richard.ell...@gmail.com wrote:
On Mar 25, 2012, at 6:51 PM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 4:18 AM, Jim Mauro james.ma...@oracle.com wrote:
If you're chasing CPU utilization, specifically %sys (time in the kernel),
I would start with
Apologies to the ZFSers, this thread really belongs elsewhere.
On Mar 25, 2012, at 10:11 PM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 11:34 AM, Richard Elling
richard.ell...@gmail.com wrote:
On Mar 25, 2012, at 6:51 PM, Aubrey Li wrote:
On Mon, Mar 26, 2012 at 4:18 AM, Jim Mauro
On Mon, Mar 26, 2012 at 12:19 PM, Richard Elling
richard.ell...@richardelling.com wrote:
Apologies to the ZFSers, this thread really belongs elsewhere.
Some of the info in it is informative for other zfs users as well though :)
Here is the output, I changed to tick-5sec and trunc(@, 5).
No.2
On Mon, Mar 26, 2012 at 1:19 PM, Richard Elling
richard.ell...@richardelling.com wrote:
Apologies to the ZFSers, this thread really belongs elsewhere.
Let me explain below:
Root documentation path of apache is in zfs, you see
it at No.3 at the above dtrace report.
The sort is in reverse
Hi,
I'm migrating a webserver(apache+php) from RHEL to solaris. During the
stress testing comparison, I found under the same session number of client
request, CPU% is ~70% on RHEL while CPU% is full on solaris.
After some investigation, zfs root lock emerges as a major doubtful point.
Firstly,
29 matches
Mail list logo
