Skip to content

memcache_hazelcast job from memcache/6.8.0

Github source: b2e2798 or master branch

Properties

memcache_hazelcast

debug

Set debug port and enable debug mode.

hazelcast

client_event_thread_count

Thread count for handling incoming event packets.

Default
5
event_thread_count

Number of event handler threads.

Default
5
executor_pool_size

The size of the executor pool to handle asyncronous requests. A value of zero will use the default of cores * 2.

Default
0
io_thread_count

Number of input and output threads.

Default
3
local_member_safe_timeout

Number of seconds to wait for local member to be safe before shutting it down.

Default
120
machines

The machines in this hazelcast cluster.

max_cache_size

The maximum amount of heap you wish to be reserved for cache entries. Once total cache size gets bigger than this then node will start LRU evicting a % from each cache. Should be less than heap_size.

Default
5.36870912e+08
max_no_heartbeat_seconds

The maximum amount of time hazelcast will fail to communicate with another node before dropping it from the cluster.

Default
60
minimum_cluster_members

Minimum number of cluster members this node can see to determine if this node is healthy in the cluster.

Default
1
operation_call_timeout

The maximum amount of time in milliseconds hazelcast will wait to get a response from a call to another node before erroring out. Making the default short cause these operations should be fast.

Default
10000
operation_generic_thread_count

Number of generic operation handler threads. -1 means CPU core count x 2.

Default
-1
operation_thread_count

Number of partition based operation handler threads. -1 means CPU core count x 2.

Default
-1
partition_count

How many partitions exist in this cluster. Effects total cluster scalability.

Default
271
percent_to_trim

When max cache size is exceeded how much should be evicted from each cache.

Default
20
port

The port this server will listen on for memcache requests.

Default
5701
receive_buffer_size

The size of the network receive buffer in KB. Hazelcast recommends this be 1M+ for a 10GB Network.

Default
1024
send_buffer_size

The size of the network send buffer in KB. Hazelcast recommends this be 1M+ for a 10GB Network.

Default
1024
trim_delay

How often in seconds should the node check to see if max cache size is exceeded.

Default
10

heap_size

The amount of Heap the cache should be configured with.

Default
1G

host

password

The memcache rest endpoint password.

port

The http port the Memcache hazelcast service will listen for REST calls on

Default
8080
username

The memcache rest endpoint username.

Default
hazelcast-user

java_opts

Any Java opts we wish to apply beyond Xmx and TMP_DIR

Default
  -Xms32m -XX:+UseG1GC -XX:MinHeapFreeRatio=5 -XX:MaxHeapFreeRatio=10 -Djava.security.egd=file:/dev/./urandom
    -Djava.awt.headless=true -Duser.timezone=UTC -Djava.net.preferIPv4Stack=true -Xss256k

memcache

max_queue_size

The maximum size of the memcache request queue before the memcache server begins to apply back pressure.

Default
100
max_value_size

The biggest a value can be in bytes.

Default
1.048576e+06
port

The port this server will listen on for memcache requests.

Default
11211
request_rate_limit

Maximum number of requests per user to allow within a 10 second window. Default is a huge value just to be safe.

Default
500000
secret_key

The shared secret used for memcache authentication.

test_cache

The cache name the test user is linked to.

test_password

The password to use when connecting with the test user.

test_user

The username that non cf clients can use to hit the cache for ping/testing purposes.

Default
test

open_file_limit

The ulimit for the maximum number of files/connections that are allowed

Default
16384

plans

The plans this server will support.

startup_timeout

The time to wait for app to startup before we kill it and let monit try again.

Default
30

Templates

Templates are rendered and placed onto corresponding instances during the deployment process. This job's templates will be placed into /var/vcap/jobs/memcache_hazelcast/ directory (learn more).

  • bin/drain (from drain.erb)
  • bin/memcache_hazelcast (from memcache_hazelcast.erb)
  • bin/post-start (from post-start.erb)
  • config/application-bosh.yml (from application-bosh.yml.erb)

Packages

Packages are compiled and placed onto corresponding instances during the deployment process. Packages will be placed into /var/vcap/packages/ directory.