Use free disk space in hosts to limit new deployments
|Assignee:||Carlos Martín||% Done:|
|Category:||Core & System|
|Target version:||Release 4.4|
Feature #1678: DS inserts clone & ln target attributes into each disk
(cherry picked from commit 4487c7cc62ecc14d1f6853858e17e52b6e21541c)
Feature #1678: Host monitorization copies DS info to each system DS. Assumes all DS are shared.
feature #1678: Datastore location sent to the IM driver. Updated probes and collector manager accordingly
Feature #1678: Add a template to HostShare, store local system DS monitorization there
Feature #1678: Make the DS monitorization more generic
System DS do not have ln and clone targets, and this
change allows monitorization to return any arbitrary
values to be used in SCHED_REQ and SCHED_RANK, like
Revert "feature #1678: Datastores report LN and CLONE target as SYSTEM, SELF or NONE"
This reverts commit 7a02a4cf9667cf105242e19fa745378d3282c56b.
feature #1678: close collectd-client.rb's output
(cherry picked from commit b66c9863707a594ea291ad0e7e93735056f37485)
feature #1678: Improve monitoring algorithm and set better oned.conf defaults
(cherry picked from commit a2b24764ace993f6ee5aa52c8af89fe0a806032f)
feature #1678: Moved update info of datastore, and minor work on monitor callback function
Feature #1678: Volatile size quotas are long long in c++, change ruby fsck & migrator to use ints instead of floats
Feature #1678: Better ds template update, clean monitorization when driver changes
Feature #1678: Do not show local system DS monitoring information in onedatastore output
feature #1678: Test functions for VM capacity uses Pool. Conditional print VM requirements for DEBUG log level
Feature #1678: The sched treats uninitialized shared system DS as non-shared ones.
Feature #1678: Simpler algorithm to distribute hosts homegeneously in collectd-client.rb
Feature #1678: Add a new MONITORING_PUSH_INTERVAL
configuration option that handles the time between push
monitoring intervals. This new parameter is passed on to the
Feature #1678: Simplify the collectd-client.rb algorithm by adding
a random sleep at the beginning.