search

bigdata4u / spymemcached

Automatically exported from code.google.com/p/spymemcached
0 stars 0 forks source link

KetamaNodeLocator does not search enough to ensure success, can see timeouts with small cluster sizes #175

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago 
It has been found that with spymemcached up through 2.5, the chances of not 
selecting another active server were 1 in 2^(number of servers).  That means 
that for a two node cluster where one node had failed, you had a one in four 
chance of not finding an active node, and receiving timeouts as a result even 
though the operation should redistribute to another node.

Thanks to Taso from Concur with help in identifying this issue.

Original issue reported on code.google.com by ingen...@gmail.com on 23 Apr 2011 at 12:36

GoogleCodeExporter commented 8 years ago

Original comment by ingen...@gmail.com on 23 Apr 2011 at 12:36