#35588: Do not try to store in cache too big values

#

Fri May 02 13:44:58 2008

KAPPA - Ticket created

Subject:

Do not try to store in cache too big values

[text/plain 681b]

memcached does not accept values bigger than 1mb. Moreover, it's
commonly not very useful to cache big values -- they are probably rare.

We here at Rambler have an option "too_big_threshold" in
Cache::Memcached which sets the size in bytes after which no storing is
even attempted. It would be nice to have the same option in Fast. The
change is very simple and I thought it would be more practical for the
author to implement it :) And I missed my XS-fu classes, too :(

Such logic cannot be elegantly implemented on the application side
because serialization is done inside Cache::Memcached.

I attach my Perl patch to the latest BRADFITZ's Cache::Memcached, it
could be useful.

Subject:

toobig.patch

[text/x-diff 1.7k]

--- Memcached.pm.orig 2007-07-17 22:40:23.000000000 +0400
+++ Memcached.pm 2008-05-02 21:43:22.000000000 +0400
@@ -25,6 +25,7 @@
bucketcount _single_sock _stime
connect_timeout cb_connect_fail
parser_class
+ too_big_threshold
};

# flag definitions
@@ -76,6 +77,7 @@
$self->{'stat_callback'} = $args->{'stat_callback'} || undef;
$self->{'readonly'} = $args->{'readonly'};
$self->{'parser_class'} = $args->{'parser_class'} || $parser_class;
+ $self->{'too_big_threshold'} = $args->{'too_big_threshold'};

# TODO: undocumented
$self->{'connect_timeout'} = $args->{'connect_timeout'} || 0.25;
@@ -460,6 +462,10 @@

my $len = length($val);

+ if ($self->{'too_big_threshold'} && $len >= $self->{'too_big_threshold'}) {
+ return 1; # positive NOP
+ }
+
if ($self->{'compress_threshold'} && $HAVE_ZLIB && $self->{'compress_enable'} &&
$len >= $self->{'compress_threshold'}) {

@@ -959,6 +965,10 @@
Values larger than this threshold will be compressed by C<set> and
decompressed by C<get>.

+Use C<too_big_threshold> to set the upper limit on size of cached data
+items. Values larger than this threshold won't be sent over the
+network. It's no use to send >1MB values anyway, they are not stored.
+
Use C<no_rehash> to disable finding a new memcached server when one
goes down. Your application may or may not need this, depending on
your expirations and key usage.
@@ -1002,6 +1012,10 @@

Sets the compression threshold. See C<new> constructor for more information.

+=item C<set_too_big_threshold>
+
+Sets the overall size threshold. See C<new> constructor for more information.
+
=item C<enable_compress>

Temporarily enable or disable compression. Has no effect if C<compress_threshold>

#

Fri May 02 14:57:18 2008

KROKI - Correspondence added

[text/plain 853b]

On Fri May 02 13:44:58 2008, KAPPA wrote:
> We here at Rambler have an option "too_big_threshold" in
> Cache::Memcached which sets the size in bytes after which no storing is
> even attempted. It would be nice to have the same option in Fast.

The request makes sense. Indeed you can't implement such threshold with
serialization and/or compression hooks in C::M::F. However, a good
question is _when_ to check the size? I'd say that it would be more
logical to test it after the compression, because that's the point when
we know what exactly is going to be stored in memcached. However in
your patch you test the size before the compression, and one may say
that this is more optimal when we know in advance that values do not
compress well. So, what you suggest? Will one test after the
compression be enough? Or should we have two thresholds?

#	Fri May 02 14:57:19 2008	RT_System - Status changed from 'new' to 'open'

#

Fri May 02 20:13:29 2008

KAPPA - Correspondence added

From:

KAPPA[...]cpan.org

[text/plain 833b]

> serialization and/or compression hooks in C::M::F. However, a good
> question is _when_ to check the size? I'd say that it would be more
> logical to test it after the compression, because that's the point when
> we know what exactly is going to be stored in memcached. However in
> your patch you test the size before the compression, and one may say
> that this is more optimal when we know in advance that values do not
> compress well. So, what you suggest? Will one test after the
> compression be enough? Or should we have two thresholds?

I think it's more practical to have one threshold and test it right
after compression. If one knows in advance that his values do not
compress well, he should switch the compression off at all.

We don't use compression, and I just forgot about it, sorry. Updated
patch attached.

[text/x-diff 1.6k]

--- Memcached.pm.orig 2007-07-17 22:40:23.000000000 +0400
+++ Memcached.pm 2008-05-03 04:09:26.000000000 +0400
@@ -25,6 +25,7 @@
bucketcount _single_sock _stime
connect_timeout cb_connect_fail
parser_class
+ too_big_threshold
};

# flag definitions
@@ -76,6 +77,7 @@
$self->{'stat_callback'} = $args->{'stat_callback'} || undef;
$self->{'readonly'} = $args->{'readonly'};
$self->{'parser_class'} = $args->{'parser_class'} || $parser_class;
+ $self->{'too_big_threshold'} = $args->{'too_big_threshold'};

# TODO: undocumented
$self->{'connect_timeout'} = $args->{'connect_timeout'} || 0.25;
@@ -474,6 +476,10 @@
}
}

+ if ($self->{'too_big_threshold'} && $len >= $self->{'too_big_threshold'}) {
+ return 1; # positive NOP
+ }
+
$exptime = int($exptime || 0);

local $SIG{'PIPE'} = "IGNORE" unless $FLAG_NOSIGNAL;
@@ -959,6 +965,10 @@
Values larger than this threshold will be compressed by C<set> and
decompressed by C<get>.

+Use C<too_big_threshold> to set the upper limit on size of cached data
+items. Values larger than this threshold won't be sent over the
+network. It's no use to send >1MB values anyway, they are not stored.
+
Use C<no_rehash> to disable finding a new memcached server when one
goes down. Your application may or may not need this, depending on
your expirations and key usage.
@@ -1002,6 +1012,10 @@

Sets the compression threshold. See C<new> constructor for more information.

+=item C<set_too_big_threshold>
+
+Sets the overall size threshold. See C<new> constructor for more information.
+
=item C<enable_compress>

Temporarily enable or disable compression. Has no effect if C<compress_threshold>

#

Sat May 03 07:34:30 2008

KROKI - Correspondence added

180 min

#	Sat May 03 07:34:32 2008	KROKI - Status changed from 'open' to 'resolved'