A few errors I was going to post about, but fixed.

[Ben]

Hi,
I am trying to use memcacheq with a few 1000 queues, with around
5-40 items each, however I am a little bit unfamiliar with BDB.

At first I received this error:
memcacheq: Logging region out of memory; you may need to increase its
size

It appears that I needed to increase my "logging region" which appears
to be a BDB option, by creating a DB_CONFIG file, and placing it in my
database's environment.

I did this by placing a file called DB_CONFIG in /data1/memcacheq

#cat DB_CONFIG
set_cachesize 0 8388608 8
set_lg_regionmax 524288
set_lg_bsize 2097152

This seemed to mitigate the particular problem I was having, but I get
another one:

I start memcacheQ:
/usr/local/bin/memcacheq -vv -r -H ./data1/memcacheq -N -L 1024 -B
128 -p 11212 -l 127.0.0.1

Then I insert items into Queues: after I insert ~2500 items, I
receive the following error message in my python script:

[WA...@1222321936.863665] mcm_server_readable():2587: timeout:
Operation now in progress: select(2) call timed out for read(2)able
fds
[IN...@1222321936.863914] mcm_get_line():1597
[WA...@1222321942.575898] mcm_server_readable():2587: timeout:
Interrupted system call: select(2) call timed out for read(2)able fds
[IN...@1222321942.576177] mcm_get_line():1597
[WA...@1222321946.384054] mcm_server_readable():2587: timeout:
Interrupted system call: select(2) call timed out for read(2)able fds
[IN...@1222321946.384323] mcm_get_line():1597

----------

On the memcacheq console I see:

fileops: write ./data1/memcacheq/./__dbq.siteq_303.0: 4096 bytes at
offset 0
fileops: write ./data1/memcacheq/siteq_304: 4096 bytes at offset 0
fileops: write ./data1/memcacheq/./__dbq.siteq_304.0: 4096 bytes at
offset 0
fileops: write ./data1/memcacheq/siteq_305: 4096 bytes at offset 0
memcacheq: memp_trickle thread: done, writing 613 dirty pages
fileops: close ./data1/memcacheq/__db.80000c5c.e2fa858
fileops: unlink ./data1/memcacheq/__db.80000c5c.e2fa858
fileops: close ./data1/memcacheq/log.0000000001
bdb_put: Too many open files
>10 NOT_STORED
add a item buffer to freelist.
<10 connection closed.
<10 new client connection
<1023 new client connection
event_add: No such file or directory
Too many open connections
<10 set siteq_509 4 0 4
alloc a item buffer from freelist.
fileops: stat ./data1/memcacheq/siteq_509
fileops: stat ./data1/memcacheq/siteq_509
fileops: stat ./data1/memcacheq/__db.80000c5e.7e3f12ab
fileops: write ./data1/memcacheq/log.0000000001: 149 bytes at offset
2730896
fileops: flush ./data1/memcacheq/log.0000000001
fileops: open ./data1/memcacheq/__db.80000c5e.7e3f12ab
fileops: close ./data1/memcacheq/__db.80000c5e.7e3f12ab
fileops: open ./data1/memcacheq/log.0000000001
memcacheq: memp_trickle thread: done, writing

-------
DBD was running out of open files.

A fix was to include ulimit -n

-Ben

I will make another post with some actual questions :-)

[Steve Chu]

Hi, ben,

memcacheq uses Berkeley DB Queue access method to storage queue
messages that is one queue per file, if you have lots of queues, then
you need to increase limit of max files that one process can open,
usually with '-c' option.

--
Steve Chu
http://stvchu.org

[Ben]

Hey steve,
thanks for your response, did you have a chance to look at my other
thread: http://groups.google.com/group/memcachedb/t/9d1bc69a20f59901


-Ben

> > [W...@1222321936.863665] mcm_server_readable():2587: timeout:

> > Operation now in progress: select(2) call timed out for read(2)able
> > fds

> > [I...@1222321936.863914] mcm_get_line():1597
> > [W...@1222321942.575898] mcm_server_readable():2587: timeout:

> > Interrupted system call: select(2) call timed out for read(2)able fds

> > [I...@1222321942.576177] mcm_get_line():1597
> > [W...@1222321946.384054] mcm_server_readable():2587: timeout:

> > Interrupted system call: select(2) call timed out for read(2)able fds

> > [I...@1222321946.384323] mcm_get_line():1597

[Steve Chu]

On Fri, Sep 26, 2008 at 10:06 AM, Ben <jami...@gmail.com> wrote:
>
> Hey steve,
> thanks for your response, did you have a chance to look at my other
> thread: http://groups.google.com/group/memcachedb/t/9d1bc69a20f59901

Could you make your problem a bit clear? So I can answer more exactly.

--
Steve Chu
http://stvchu.org