Hey, Instead of setting a date, use an expiration of "3600 * 6" to mean six hours? It seems that's what you're attempting to do but that routine might not be working out for you. Don't add "3600 * 6" to anything, just set the expiration to be the result of that.
-Dormando On Tue, 1 Sep 2009, [email protected] wrote: > > Hello folk! > > Currently, we use memcached as caching solution. > We provide a link to 3 servers, a server-2x the Memcache daemon is > started (2 different ports (depending 4GB) > > 192.168.0.1:11211 (4 GB) > 192.168.0.1:11212 (4 GB) > 192.168.0.2:11211 (700 MB) > 192.168.0.3:11211 (700 MB) > > -> Php.ini: memcache.hash_strategy consistent = <-- > > $o_memcache = new Memcache (); > $o_memcache->addServer ([SERVER], [PORT]); > > whenever a request comes the "connect" is no longer needed because > Memcache established the connection automatically. > We also tried $o_memcache=connect([SERVER], [PORT]); - same problem. > > About $o_memcache->get([KEY]), we check to see if the KEY is already > set (KEY = UNIQUE!). > In case of error returns GET "false". > When TRUE, the content of the KEY will be shown - otherwise with > $o_memcache->set([KEY], [CONTENT], MEMCACHE_COMPRESSED, mktime (date > ( "H") +6)); will be saved for 6 hours. > > You should guess the KEY will be saved for the next 6 hours, but in > this case, it will not. > After about 2-3 minutes the KEY returns with GET "FALSE" again. > > > We also tried this: > - Memcache connection with addServer, connect and pconnect > - 1 server as connection, multiple server (like listed above - as > pool) > > > Server stats: > evictions = 0 > cur_connections: ~50-100 > > > Server settings: > > Apache 2.11 > PHP 5.2.8 > > Memcache-Info by phpinfo(); > Active persistent connections 4 > Version 2.2.5 > Revision $Revision: 1.111 $ > memcache.allow_failover 1 1 > memcache.chunk_size 8192 8192 > memcache.default_port 11211 11211 > memcache.default_timeout_ms 1000 1000 > memcache.hash_function crc32 crc32 > memcache.hash_strategy consistent consistent > memcache.max_failover_attempts 20 20 > > > Does anybody know this issue or solution(s) to fix this problem? > > Thanks for help. > > Jean Michel > > PS: Also used memcached (with a "d"). Same Problem... >
