+1
One concern that came across my mind was accidentally performing
incr/decr on a wrong key (with a valid value for something else) but
this is probably unlikely to happen... Even if it did happen, its the
web developer's fault anyway ;)
Toru
On Tue, Sep 2, 2008 at 7:39 AM, dormando <[EMAIL PR
inary protocol, if an item didn't exist it will
make an initial one for you.
Personally I like what memcached does for incr/decr with the binary
protocol but wouldn't this become a potential trap for those that are
used to having NOT_FOUND sent back from memcached?
Toru
On Tue, Sep 2, 2008 at
Hi!
> We are working on creating a "storage interface" to memcached, so that you
> can create you own back-end. This sounds like a pretty easy task to
> implement in the prototype we have
Indeed, with the pluggable storage engine that is going on, whats
been debated in this thread would be
Hi!
> I've updated the wiki. Does anyone else have proposed areas to focus on in
> this hackathon? Personally, I'd like to see some time put into some of the
> client work. Hopefully some of the client maintainers can make it?
I've been thinking about the client libraries too. The binary tree
Hi!
Letting everyone know that I did a presentation on the binary protocol few
nights ago and that I've shared the slides on:
http://www.slideshare.net/tmaesaka/memcached-binary-protocol-in-a-nutshell-presentation/
In brief the presentation covers the following:
- Why Binary?
- How we can benefi
Hi!
FYI, from version 1.3 multi-thread build is going to be default (it
already is in the development branch).
In addition to Todd's reply, if you actually test the performance
between a single thread build and a multi thread build, you will
notice that the performance difference is not so signi
Hi,
About the opaque issue in the stats implementation, as much as I love
the 0x746f7275 idea (it would be funny)
Trond came up with a very flexible solution that looks like the way to
go (this happened at like 1am).
So, as far as the binary protocol spec goes, the server should really
send back
Oh, and I've shared the photos I've taken last night on Flickr too:
http://www.flickr.com/photos/tmaesaka/sets/72157608050620544/
On Thu, Oct 16, 2008 at 12:53 AM, Toru Maesaka <[EMAIL PROTECTED]> wrote:
> Hi,
>
> About the opaque issue in the stats implementation
Hi guys,
I fixed the stats opaque issue pointed out by Dustin's test case at
the hackathon and just pushed it to github:
http://github.com/tmaesaka/memcached/commit/86159eb7b5c611d25a33eb1cb6c750c806350912
I also refactored 'server_stats()' which is a function that returns
server specific stats
Hi!
I'm not sure if we could call it a waste since CAS is a important
feature but sure, it would consume a lot of memory in usecases with
lots and lots of small objects as you mentioned. I guess the ideal
thing is to only consume that 8 bytes when CAS is actually required
(e.g. when we have req_c
the ``move'' commit
>> empty (and I removed empty commits). This gives us a much better
>> history with the following authors (again, some google searches, so if
>> some email addresses are wrong, I apologize):
>>
>> 158 Brad Fitzpatrick <[EMAIL PROTECTED]
Hi,
If you MUST, go into slabs.c and grep for the constant, 'POWER_BLOCK'.
Cheers,
Toru
On Tue, Dec 9, 2008 at 2:37 PM, bsergean <[EMAIL PROTECTED]> wrote:
>
> Hi,
>
> http://www.socialtext.net/memcached/index.cgi?faq#what_are_some_limits_in_memcached_i_might_hit
> """
> Stored data cannot exce
Hi!
I've quickly glanced through the repo and even though I think the idea
of the flat allocator is neat and how it uses mmap (/me loves mmap), I
think the flat allocator itself should belong outside the codebase.
So, what I'm trying to say is, shouldn't this be a separate storage
engine?
Trond
oops, clicked send before I finished the email :(
ASFAIK, Trond and Matt has been experimenting quite a bit with what
I've mentioned above.
Thanks,
Toru
On Sat, Dec 20, 2008 at 9:42 PM, Toru Maesaka wrote:
> Hi!
>
> I've quickly glanced through the repo and even though I
Hi!
This is awesome! you've put a lot of useful content up there :) ++
Toru
On Mon, Dec 29, 2008 at 4:21 PM, Joonas Hämäläinen
wrote:
>
> On Mon, Dec 29, 2008 at 7:55 AM, dormando wrote:
>> I went over the wiki with some steel wool and a blowtorch:
>> http://code.google.com/p/memcached/
>>
>>
Hi!
I personally think that the behavior of not responding at all is okay
since this is for complementing the no-reply mode but you do have
a point about the relationship with getq...
Here's a thought, could we separate the set commands?
Cheers,
Toru
On Mon, Jan 26, 2009 at 1:47 PM, Dustin wr
the binary protocol? where's dormando ;)
Cheers,
Toru
On Mon, Jan 26, 2009 at 3:52 PM, Dustin wrote:
>
>
> On Jan 25, 10:13 pm, Toru Maesaka wrote:
>
>> I personally think that the behavior of not responding at all is okay
>> since this is for complementing the no-r
As for performance, I found that there is better throughput as
concurrency grows when I tested the 1.3 tree a while back.
I haven't actually profiled the internals (really should with dtrace..),
but I suspect that it is the performance gain in the parser that
caused this gain (as we would expect)
Hi,
'stats maps' imho, isn't so useful since we can get the same information
with commands like ldd(1). This doesn't need to be in the server since
low layer folks that are interested in this information would know what
to do to obtain this information.
'stats malloc', would be useful for debugg
still convinced a full dump command will be abused more than it's
> worth, but I forget where we left off last on that debate.
>
> -Dormando
>
> On Wed, 28 Jan 2009, Toru Maesaka wrote:
>
>>
>> Hi,
>>
>> 'stats maps' imho, isn't so use
G'day
There's an inconsistency between the ascii protocol documentation and
the actual server implementation regarding stats. Specifically, I'm talking
about rusage_user and rusage_system entries.
So, in docs/protocol.txt it is stated that:
" '32u:32u' means two 32-but unsigned integers separat
Cool. I was going to start experimenting with concurrency strategies
on that tree so this info helps.
On Fri, Mar 6, 2009 at 9:36 AM, Dustin wrote:
>
>
> To anyone who pays attention to my server repo: I've set my master
> to what was rewritten-bin because it's been a bit confusing for me and
>
Hi!
I was going to reply to Dustin's thread but noticed we haven't talked
about development related goodness in a while so I'm starting a new
thread.
So, to begin with I'm going to describe what I'm going to experiment
to get memcached scaling even further in modern hardware. As most of
you know
would work out... I'll look into it and report it to you guys
again :)
Cheers,
Toru
On Fri, Mar 6, 2009 at 2:04 PM, Dustin wrote:
>
>
> On Mar 5, 7:11 pm, Toru Maesaka wrote:
>
>
>> As for the global stats_lock, I think its safe to assume that most of
>> the operat
Forgot to mention in my previous email that global operations like
stats can be done by locking all engine instances and aggregate the
result on demand.
On Fri, Mar 6, 2009 at 6:37 PM, Toru Maesaka wrote:
> Hi Dustin, Ben and all
>
>> Trond's already done that per-thread work
to go. Damn, really wanted to
achieve minimal locks internally...
Trond: Hopefully talk to you on IRC this afternoon :)
On Fri, Mar 6, 2009 at 8:39 PM, Trond Norbye wrote:
>
> Dustin wrote:
>>
>> On Mar 5, 7:11 pm, Toru Maesaka wrote:
>>
>>
>>
>>>
>&g
Report feedback to the list
> or file bugs as you find them.
>
> * Mailing List: [http://groups.google.com/group/memcached]
> * Issue Tracker: [http://code.google.com/p/memcached/issues/list]
> * IRC: #memcached on freenode
>
> 5 Contributors
> ~~~
>
>
spool four threads
by default but you can change this with the -t option.
Cheers!
Toru
On Thu, Mar 12, 2009 at 3:09 PM, Toru Maesaka wrote:
> Awesome! thanks big time for getting this out.
>
> One thing I'd like to add to Dustin's message is the space
> efficiency option
Personally I find this weird but I see your motivation so I have no objections.
If it's going to look like we're going to be adding more options in
the future, maybe its worth considering moving out the getopt call
from the main (wrap it in get_arguments() or something).
On Tue, Mar 17, 2009 at
Heh, it's on my to-review list. Lemmie get back to you within the next 24hrs ;)
On Wed, Mar 18, 2009 at 2:39 AM, Dustin wrote:
>
>
> This just scrolled off the list page with no commentary. I'm going
> to assume it's because everybody think it's completely awesome and/or
> nobody would dare qu
test passes both on Mac OS X and
Ubuntu 8.04 64bit. It's really nice how we have a dedicated file for
utility functions now, things like this are always handy ;)
Cheers,
Toru
On Wed, Mar 18, 2009 at 3:46 PM, Dustin wrote:
>
>
> On Mar 17, 7:05 pm, Toru Maesaka wrote:
>>
Hi!
Just looked through the commits. Yeah, it does make sense to add other
strto* functions that we're using in the core like you did in that
tree.
The test passes in my environment and the tiny issue I pointed out
earlier is fixed so I have nothing to complain about :)
Cheers,
Toru
On Thu, Ma
Hi!
This is not what memcached should do.
If a particular item doesn't exist, the daemon should not have to do
anything extra. If you get a miss, it's a miss. This is a golden rule
that should not be complicated nor broken.
If you want to populate the missing item, do it after you get a miss
fr
Hi!
So we've been testing memcached-1.3.2 beta at mixi.jp for
a week and we found an issue in the stats code.
The problem is that the server segfaults if we issue the
"stats slabs" command to a daemon that has lots of slab
classes (approximately 40).
The cause is simply not allocating enough me
You're creating a hash value with the MD5 algorithm. Not sure why it's
prefixing with 'query' since that's not really namespacing anything
(afraid of collision perhaps?).
http://php.net/md5
Toru
On Thu, Apr 16, 2009 at 12:28 PM, Paras wrote:
>
> I found a sample code online which showed me how
Hi!
I would usually stay out of things like this but could you
please be less offensive? Saying things like this to someone
that shared something nice to the public is beyond rude.
This is clearly bad community spirit.
How would you feel if I (or the community) torched the new
code that you're w
cachedump is a debugging functionality for developers, meaning it's
not an officially supported command. Thats why its undocumented.
If you see a command thats not documented in the spec, it's a client
specific feature that is implemented on the client side.
Toru
On May 8, 2009, at 9:54
What exactly is the error that you're getting? we can help you more if
you paste the error that you're getting.
Are you suspecting that memcached is behaving oddly due to a count
mismatch from your application? If so, it could be possible that
you're doing something funky in your script.
It woul
Hi!
Apologies for the late reply.
Things you've mentioned are correct and we've talked about
SMP performance issues of the daemon for a while at hackathons,
events, IRC and etc. As Dustin had mentioned, some experimental
work has been done and hopefully things will look better in future
releases
Hi!
The default growth factor used in memcached is 1.25 which you can
control with the '-f' option (take a look at the help menu with '-h').
The lower limit of a slab class is 88 bytes and the upper limit is 1MB.
The ideal growth factor depends on your use-case and if you're
interested in playi
G'day!
So I just had a look at the patch and from reading Adrian's reply, I
can understand why such a feature would be handy from hosting
perspective.
The concern is, I think that the majority of users won't need this
feature. That said, it doesn't change the fact that there's a need for
it so h
tion them.
>
> Is there an existing way to use the memcached "gets" and "cas"
> commands from Perl, without writing my own module to interface to
> memcached directly?
>
> -Bennett
>
>
--
Toru Maesaka
Awesome! I think this is a handy change to get potential bugs
identified and fixed as soon as possible.
On Aug 3, 2009, at 5:50 PM, dormando wrote:
Yo,
Updates to the issue tracker should now all hit the mailing list. The
traffic over there is relatively low, but only a small handful of
Hi!
As of version 1.3 multi-thread build is default in memcached. So you
don't need to do that with the latest 1.4 series.
What version are you trying to build?
Cheers,
Toru
On Sep 22, 2009, at 11:37 PM, bcarey wrote:
Hi everyone,
I'm trying to run the configure script with --enable-t
+1
Warnings are good.
On Tue, Oct 6, 2009 at 4:08 PM, dormando wrote:
>
> Yo,
>
> I'm debating a small patch for 1.4.2 that would add (more) commandline
> warnings for memcached. You'd of course only see most of them if you fire
> it up without the -d option, but hopefully in the course of test
Hi folks!
>From chasing the commit log and reading Trond's blog entry, I noticed
that we're throwing in SASL support to memcached.
I guess this is to make it friendlier to deploy memcached on an
untrusted network (e.g. Amazon's EC2) but I wanted to hear what the
actual deciding factor was. You k
eople that has very limited
freedom over the infrastructure?
Cheers,
Toru
On Tue, Oct 27, 2009 at 1:09 AM, Dustin wrote:
>
>
> On Oct 26, 1:33 am, Toru Maesaka wrote:
>
>> From chasing the commit log and reading Trond's blog entry, I noticed
>> that we're th
Awesome!
On Sat, Nov 7, 2009 at 4:33 PM, dormando wrote:
>
>
> Coinciding with the release of 1.4.3 tomorrow, we will be retiring the old
> and tired http://danga.com/memcached homepage (archiving it, I guess). It
> will redirect to http://memcached.org, which will have this new site:
>
> http:/
dicated to one label/auth or
>> something like that, right?
>
> It may be possible. However, if the usage of slabs are not balanced,
> it may remove cached object, although we have unallocated cache in
> the slab owned by other label/auth. right?
>
> Thanks,
> --
> OSS Platform Development Division, NEC
> KaiGai Kohei
>
--
Toru Maesaka
t; It'd be incredibly inefficient on memory, compared to keeping the number
>>>> of users down or even running multiple instances.
>>>
>>> Only if you were trying to go all the way down to the item level. It's
>>> possible to have groups of slabs that are dedicated to one label/auth or
>>> something like that, right?
>>
>> It may be possible. However, if the usage of slabs are not balanced,
>> it may remove cached object, although we have unallocated cache in
>> the slab owned by other label/auth. right?
>>
>> Thanks,
>> --
>> OSS Platform Development Division, NEC
>> KaiGai Kohei
>>
>
>
>
> --
> Toru Maesaka
>
--
Toru Maesaka
ems to me we will
> face a tradeoff between persistent storage and access controls.
No. It's not stackable. If you want to do this you'd have to create an
abstraction layer within the storage engine. There were discussions on
this a while back though (I forget if it was online or offl
Yeah it would be neat to see a separate "Slabber with Access Control"
engine.
On Jan 7, 2010, at 5:50 PM, dormando wrote:
Are you suggesting that applications has to handle the scramble
buffer
correctly for each accesses? It seems to me we can obtain
credential of
the client using SASL a
for not realizing this earlier... Good luck on
hacking your Access Control engine module :)
- Toru
P.S. It might save your time if you look at this first:
http://github.com/trondn/memcached/blob/engine/doc/engine-interface.xml
2010/1/7 KaiGai Kohei :
> (2010/01/07 17:55), Toru Maesaka
t only memcached.
>>
>> How much length do you expect for the security property of an item?
>> I plan to add a "void *" pointer which indicates security_id object
>> managed in libselinux. The security_id structure is a pair of
>> security context (cstring) and reference counter. The number of
>> security context is proportional to the number of users, not # of items,
>> it does not cousume local memory so much.
>>
>>>> * The deployment of some hooks makes hard to decide what permission
>>>> should be applied. For example, process_bin_delete() calls "get"
>>>> method to confirm the object to be removed is existing.
>>>> But the "get" method is only chance for the engine module to know
>>>> a certain item is accessed in reader context.
>>>> Of course, if we consider "delete" is a combination of "read" and
>>>> "delete", it will be resolved. But it will damage the flexibility
>>>> of security design.
>>>
>>> This is a good time to be talking about this, as the engine API
>>> isn't dogma yet.
>>
>> Now I lean toward the security purpose framework which provides
>> meaningful hooks for access controls, and capability to assign
>> a security property on the session and items.
>>
>> Is it too early?
>>
>> Thanks,
>
>
> --
> OSS Platform Development Division, NEC
> KaiGai Kohei
>
--
Toru Maesaka
Hi,
$ memcached -h | grep socket
-s UNIX socket path to listen on (disables network support)
-a access mask for UNIX socket, in octal (default: 0700)
- Toru
On Fri, Feb 12, 2010 at 8:52 AM, Brian Hawkins wrote:
> I've searched the internet and cannot find a conclusive answer to whethe
Hi!
The replies in the thread you supplied seems to answer your question.
Toru
On Jun 23, 2010, at 4:09 PM, MuFei wrote:
Hi,
I've successfully installed Memcached and it seems to be running. But
it has no cached data and it does not cache any data. after some
googling I found some that ther
+1
Regardless of using pthread's conditional variable or not, the idea is
consistent.
It's truly awesome that a test script is provided. Great work Mala! and
Twittervese for investigating the problem.
-Toru
-Original Message-
From: dormando
Sender: memcached@googlegroups.com
Date: Tu
You don't have to see anyone to present your work.
Good start would be to fork memcached on GitHub, create your own branch and
send a pull request to upstream with the patch & description of what you're
trying to accomplish.
https://github.com/memcached/memcached
More info can be found here:
htt
You need to define big. How big are these values?
As you hinted, SERVER_MAX_VALUE_LENGTH isn't related to memcached itself.
>From a quick glimpse of python3-memcache code, it seems to silently fail if
the value length is greater than SERVER_MAX_VALUE_LENGTH. See:
https://github.com/pilif/python-
Glad to hear you found the problem!
I don't have 1.4.4 on me but the current release (1.4.15) doesn't crash
even if you give it a value great than 5MB. Maybe it's a good time for you
to upgrade. If it still behaves oddly then the problem is in your client.
On Tue, Feb 12, 2013 at 4:55 PM, Patric
Getting straight to the point -- this would be a good start.
https://github.com/memcached/memcached/blob/master/items.c#L88-L216
On Fri, May 10, 2013 at 6:27 AM, Nicola Licheri wrote:
> Hi.
> For my thesis, I'm working on a fork of Memcached, a C++ reimplementation
> without libevent.
> To do t
Oh bummer! I definitely would have flown in from the land of the rising sun
if there was earlier notice.
Although... I can still make it if I fly out this afternoon...
:dr-evil-face:
On Tue, May 21, 2013 at 7:31 AM, Brad Fitzpatrick wrote:
> Hey,
>
> memcached is 10 years old this Wednesday th
Pointing out the obvious but why don't you start by restricting port 11211
(or whatever you're using). You could then strategically "move" your
memcached instance(s) to host(s) that are not Internet accessible.
On Thu, Jun 5, 2014 at 3:30 PM, Rambilas Varma wrote:
> Hi Team,
>
> I am getting "u
Hi there,
Request latency can be caused by so many different factors. I don't think
any of us can provide definitive feedback based on the information you
presented, but generally speaking, ms to s latency is very uncommon.
It would be very helpful if you could gather evidence of your network
lat
64 matches
Mail list logo