[support] Problem with Drupals built in cache
Jamie Holly
hovercrafter at earthlink.net
Thu Jan 20 17:06:25 UTC 2011
You're using the regular memcache module. I've never used that one.
Maybe someone who has would have an idea. Personally, I use cacherouter
since it offers more features and flexibility.
Jamie Holly
http://www.intoxination.net
http://www.hollyit.net
On 1/20/2011 10:32 AM, Margusja wrote:
> In error log there is no problems recorded.
>
> tnx for link. I'll check it out. Maybe it helps me closer to resolve my
> problem.
>
> config lines from settings.php
> $conf['cache_inc'] = './sites/all/modules/memcache/memcache.inc';
> $conf['memcache_key_prefix'] = 'seb_default';
>
> Tervitades, Margus (Margusja) Roo
> +372 51 48 780
> http://margus.roo.ee
> msn: margusja at kodila.ee
> skype: margusja
>
>
> On 1/20/11 5:14 PM, Jamie Holly wrote:
> > You got a lot of misses in there. Have you checked your error logs to
> > see if anything is going on? One thing that is possible is that you are
> > hitting the 1mb limit for objects in memcache with the variable cache.
> > This thread has some details on it:
> >
> > http://groups.drupal.org/node/116764
> >
> > Also if you could post your memcache configuration from settings.php to
> > see if anything is going on in there.
> >
> > Jamie Holly
> > http://www.intoxination.net
> > http://www.hollyit.net
> >
> >
> > On 1/20/2011 10:01 AM, Margusja wrote:
> >> Tnx for hints.
> >>
> >> I enabled memcache in my PHP conf. And I think the module is loaded.
> >> [16:50:57 root at arendus ~]# php -i | grep memcache
> >> memcache
> >> memcache support => enabled
> >> memcache.allow_failover => 1 => 1
> >> memcache.chunk_size => 32768 => 32768
> >> memcache.compress_threshold => 20000 => 20000
> >> memcache.default_port => 11211 => 11211
> >> memcache.hash_function => crc32 => crc32
> >> memcache.hash_strategy => consistent => consistent
> >> memcache.lock_timeout => 15 => 15
> >> memcache.max_failover_attempts => 20 => 20
> >> memcache.protocol => ascii => ascii
> >> memcache.redundancy => 1 => 1
> >> memcache.session_redundancy => 2 => 2
> >> Registered save handlers => files user memcache
> >>
> >> [16:49:22 root at arendus ~]# echo stats | nc localhost 11211
> >> STAT pid 30211
> >> STAT uptime 23950
> >> STAT time 1295535057
> >> STAT version 1.2.8
> >> STAT pointer_size 32
> >> STAT rusage_user 0.130980
> >> STAT rusage_system 0.311952
> >> STAT curr_items 473
> >> STAT total_items 1763
> >> STAT bytes 587316
> >> STAT curr_connections 10
> >> STAT total_connections 242
> >> STAT connection_structures 15
> >> STAT cmd_flush 0
> >> STAT cmd_get 9007
> >> STAT cmd_set 1763
> >> STAT get_hits 1025
> >> STAT get_misses 7982
> >> STAT evictions 0
> >> STAT bytes_read 11112904
> >> STAT bytes_written 8412213
> >> STAT limit_maxbytes 67108864
> >> STAT threads 5
> >> STAT accepting_conns 1
> >> STAT listen_disabled_num 0
> >> END
> >>
> >> Still in mysqld log:
> >> 110120 16:53:15 4102 Connect user at localhost on db
> >> 4102 Query SET NAMES "utf8"
> >> 4102 Query SELECT 1 FROM access WHERE type = 'host'
> >> AND 'IP' LIKE mask AND status = 0 LIMIT 0, 1
> >> 4102 Query INSERT INTO semaphore (name, value, expire)
> >> VALUES ('variable_cache_regenerate',
> >> '14419249754d384c5bd38895.47612777', 1295535225.8664)
> >> 4102 Query SELECT * FROM variable
> >> 4102 Query DELETE FROM semaphore WHERE name =
> >> 'variable_cache_regenerate' AND value = '14419249754d384c5bd38895.47612777'
> >> 4102 Query SELECT name, filename, throttle FROM system
> >> WHERE type = 'module' AND status = 1 AND bootstrap = 1 ORDER BY weight
> >> ASC, filename ASC
> >> 4102 Query SELECT * FROM languages ORDER BY weight
> >> ASC, name ASC
> >> 4102 Query SELECT dst FROM url_alias LIMIT 0, 1
> >>
> >>
> >>
> >> Tervitades, Margus (Margusja) Roo
> >> +372 51 48 780
> >> http://margus.roo.ee
> >> msn: margusja at kodila.ee
> >> skype: margusja
> >>
> >>
> >> On 1/20/11 4:03 PM, Jamie Holly wrote:
> >>> Then I would check phpinfo to make sure that the memcache extension is
> >>> loading into PHP (you can do this from admin -> reports -> status and
> >>> then clicking on your PHP version number).
> >>>
> >>> If that's showing up ok, you can run a command from the terminal to make
> >>> sure things are being written/read from memcache:
> >>>
> >>> echo stats | nc {memcache address} {memcache port}
> >>>
> >>> Your main things to look at are the hits, misses, bytes and connections.
> >>>
> >>> There are also some versions of memcache + the PHP extension that don't
> >>> work well together. This can be a hit and miss thing finding them. I
> >>> know there was a combination in the CentOS repos a year or so ago that
> >>> caused problems. Getting the latest version of the PHP extension and
> >>> building it from source would fix that.
> >>>
> >>> Jamie Holly
> >>> http://www.intoxination.net
> >>> http://www.hollyit.net
> >>>
> >>>
> >>> On 1/20/2011 8:43 AM, Margusja wrote:
> >>>> Thx for a replay. I checked again my memcache and I think it works fine.
> >>>>
> >>>> [15:40:16 root at arendus ~]# telnet localhost 11211
> >>>> Trying 127.0.0.1...
> >>>> Connected to localhost.
> >>>> Escape character is '^]'.
> >>>> stats
> >>>> STAT pid 30211
> >>>> STAT uptime 19724
> >>>> STAT time 1295530831
> >>>> STAT version 1.2.8
> >>>> STAT pointer_size 32
> >>>> STAT rusage_user 0.094985
> >>>> STAT rusage_system 0.230964
> >>>> STAT curr_items 67
> >>>> STAT total_items 1089
> >>>> STAT bytes 168012
> >>>> STAT curr_connections 11
> >>>> STAT total_connections 188
> >>>> STAT connection_structures 14
> >>>> STAT cmd_flush 0
> >>>> STAT cmd_get 6580
> >>>> STAT cmd_set 1089
> >>>> STAT get_hits 711
> >>>> STAT get_misses 5869
> >>>> STAT evictions 0
> >>>> STAT bytes_read 9254362
> >>>> STAT bytes_written 5742821
> >>>> STAT limit_maxbytes 67108864
> >>>> STAT threads 5
> >>>> STAT accepting_conns 1
> >>>> STAT listen_disabled_num 0
> >>>> END
> >>>>
> >>>> rows from settings.php
> >>>> if(extension_loaded('memcache')){
> >>>> $conf['cache_inc'] = './sites/all/modules/memcache/memcache.inc';
> >>>> $conf['memcache_key_prefix'] = 'xxx';
> >>>> }
> >>>>
> >>>> Tervitades, Margus (Margusja) Roo
> >>>> +372 51 48 780
> >>>> http://margus.roo.ee
> >>>> msn: margusja at kodila.ee
> >>>> skype: margusja
> >>>>
> >>>>
> >>>> On 1/20/11 2:45 PM, Jamie Holly wrote:
> >>>>> Drupal is still going to check the access table to see if you have any
> >>>>> blocks in there for the current user, even if the table is empty.
> >>>>>
> >>>>> The semaphore you are seeing is to prevent a cache stampede. You said
> >>>>> that once you disabled memcache, everything worked fine. It looks like
> >>>>> you have a problem with either your memcache server, the PHP memcache
> >>>>> extension or your settings in Drupal for memcache. I would check the
> >>>>> server first, since that's the easier. You can just do a telnet to check
> >>>>> that out. From shell:
> >>>>>
> >>>>> telnet (server address) (server port)
> >>>>>
> >>>>> See if it connects OK.
> >>>>>
> >>>>> As far as Pressflow+Memcache, yeah they work great together. I run that
> >>>>> combination for a couple of clients, including one that has seen over
> >>>>> 150,000 page an hour, and never had a problem.
> >>>>>
> >>>>> Jamie Holly
> >>>>> http://www.intoxination.net
> >>>>> http://www.hollyit.net
> >>>>>
> >>>>>
> >>>>> On 1/20/2011 4:12 AM, Margusja wrote:
> >>>>>> If I disable memcache support in PHP then in the problematic server I
> >>>>>> see that there no problem anymore. So can memcache and pressflow cache
> >>>>>> work parallel ?
> >>>>>>
> >>>>>> Tervitades, Margus (Margusja) Roo
> >>>>>> +372 51 48 780
> >>>>>> http://margus.roo.ee
> >>>>>> msn: margusja at kodila.ee
> >>>>>> skype: margusja
> >>>>>>
> >>>>>>
> >>>>>> On 1/20/11 10:09 AM, Margusja wrote:
> >>>>>>> Hi
> >>>>>>>
> >>>>>>> I have two separate Drupal instances (Pressflow 6.19) In first one the
> >>>>>>> local cache works fine. Sql log:
> >>>>>>> 326 Query SET NAMES "utf8"
> >>>>>>> 326 Query SELECT 1 FROM access WHERE type = 'host' AND
> >>>>>>> '195.50.208.170' LIKE mask AND status = 0 LIMIT 0, 1
> >>>>>>> 326 Query SELECT data, created, headers, expire,
> >>>>>>> serialized FROM cache WHERE cid = 'variables'
> >>>>>>> 326 Query SELECT data, created, headers, expire,
> >>>>>>> serialized FROM cache_page WHERE cid = 'http://IP/path/to/node'
> >>>>>>> 326 Query SELECT name, filename, throttle FROM system
> >>>>>>> WHERE type = 'module' AND status = 1 AND bootstrap = 1 ORDER BY weight
> >>>>>>> ASC, filename ASC
> >>>>>>> 326 Query SELECT name, filename, throttle FROM system
> >>>>>>> WHERE type = 'module' AND status = 1 AND bootstrap = 1 ORDER BY weight
> >>>>>>> ASC, filename ASC
> >>>>>>> 326 Quit
> >>>>>>>
> >>>>>>> But If I request the same node in the second machine. The code is same,
> >>>>>>> then the Sql log is:
> >>>>>>> 122 Quit
> >>>>>>> 123 Query SET NAMES "utf8"
> >>>>>>> 123 Query SELECT 1 FROM access WHERE type = 'host'
> >>>>>>> AND 'IP' LIKE mask AND status = 0 LIMIT 0, 1
> >>>>>>> 123 Query INSERT INTO semaphore (name, value, expire)
> >>>>>>> VALUES ('variable_cache_regenerate',
> >>>>>>> '17865562914d37e5d81acd80.47344784', 1295508982.1098)
> >>>>>>> 123 Query SELECT * FROM variable
> >>>>>>> 123 Query DELETE FROM semaphore WHERE name =
> >>>>>>> 'variable_cache_regenerate' AND value = '17865562914d37e5d81acd80.47344784'
> >>>>>>> 123 Query SELECT name, filename, throttle FROM system
> >>>>>>> WHERE type = 'module' AND status = 1 AND bootstrap = 1 ORDER BY weight
> >>>>>>> ASC, filename ASC
> >>>>>>> 123 Query SELECT * FROM languages ORDER BY weight
> >>>>>>> ASC, name ASC
> >>>>>>> 123 Query SELECT dst FROM url_alias LIMIT 0, 1
> >>>>>>> 123 Query SELECT src FROM url_alias WHERE dst =
> >>>>>>> 'hoiused-ja-investeerimine' AND language IN('et', '') ORDER BY language
> >>>>>>> DESC, pid DESC
> >>>>>>> 123 Query SELECT name, filename, throttle FROM system
> >>>>>>> WHERE type = 'module' AND status = 1 ORDER BY weight ASC, filename ASC
> >>>>>>> 123 Query INSERT INTO semaphore (name, value, expire)
> >>>>>>> VALUES ('locale_cache_et', '17865562914d37e5d81acd80.47344784',
> >>>>>>> 1295508982.4035)
> >>>>>>> 123 Query SELECT s.source, t.translation, t.language
> >>>>>>> FROM locales_source s LEFT JOIN locales_target t ON s.lid = t.lid AND
> >>>>>>> t.language = 'et' WHERE s.textgroup = 'default' AND s.version = '6.19'
> >>>>>>> AND LENGTH(s.source)< 75
> >>>>>>> 123 Query DELETE FROM semaphore WHERE name =
> >>>>>>> 'locale_cache_et' AND value = '17865562914d37e5d81acd80.47344784'
> >>>>>>> 123 Query SELECT s.lid, t.translation, s.version FROM
> >>>>>>> locales_source s LEFT JOIN locales_target t ON s.lid = t.lid AND
> >>>>>>> t.language = 'et' WHERE s.source = 'Stores details about batches
> >>>>>>> (processes that run in multiple HTTP requests).' AND s.textgroup =
> >>>>>>> 'default'
> >>>>>>> ....
> >>>>>>> ....
> >>>>>>> loads of SQLs
> >>>>>>>
> >>>>>>> The main question why does the last one ask something from semaphore? In
> >>>>>>> both servers access table is empty. At first I thought the condition
> >>>>>>> came from that SQL: SELECT 1 FROM access WHERE type = 'host' AND 'IP'
> >>>>>>> LIKE mask AND status = 0 LIMIT 0, 1 but as I said in both servers access
> >>>>>>> table is empty.
> >>>>>>>
> >>>>>>> In both solutions caches are enabled.
> >>>>>>>
> >>>>>>> Any hints?
> >>>>>>>
More information about the support
mailing list