Markus,
My configuration is as follows...
filterCache
class=solr.FastLRUCache
size=16384
initialSize=16384
autowarmCount=16384/
queryResultCache
class=solr.LRUCache
size=16384
initialSize=16384
autowarmCount=16384/
documentCache
class=solr.LRUCache
size=16384
initialSize=16384
autowarmCount=16384/
...
useColdSearcherfalse/useColdSearcher
maxWarmingSearchers2/maxWarmingSearchers
...
useCompoundFilefalse/useCompoundFile
ramBufferSizeMB64/ramBufferSizeMB
mergeFactor10/mergeFactor
unlockOnStartupfalse/unlockOnStartup
reopenReaderstrue/reopenReaders
No cache warming queries and our machines have 8g of memory in them with
about 5120m of ram dedicated to so Solr. When our index is around 10-11g
in size everything runs smoothly. At around 20g+ it just falls apart.
Can you (or anyone) provide some suggestions? Thanks
On 12/12/10 1:11 PM, Markus Jelsma wrote:
There can be numerous explanations such as your configuration (cache warm
queries, merge factor, replication events etc) but also I/O having trouble
flushing everything to disk. It could also be a memory problem, the OS might
start swapping if you allocate too much RAM to the JVM leaving little for the
OS to work with.
You need to provide more details.
After replicating an index of around 20g my slaves experience very high
load (50+!!)
Is there anything I can do to alleviate this problem? Would solr cloud
be of any help?
thanks
