Comment #4 on issue 748 by [email protected]: verify scaling of inotify in
luxid
http://code.google.com/p/ganeti/issues/detail?id=748
commit 3ebed1a18705d0977ed29ad0a97405ce99fa3e25
Author: Klaus Aehlig <[email protected]>
Date: Wed Jul 2 15:58:07 2014 +0200
Describe the new scheduling parameter in NEWS
We have added a new scheduling feature, so describe it in NEWS.
By default, it has no effect (the default for --max-tracked-jobs
is higher than the default of --max-running-jobs), but it explains
how luxid will keep reactive if --max-running-jobs is increased.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit b1365e8f277583d386cba71638eb9ccc2db5564d
Author: Klaus Aehlig <[email protected]>
Date: Wed Jul 2 13:40:44 2014 +0200
Make the scheduler use the max_tracked_jobs config parameter
Use the run-time configuration to decide on whether to attach an
inotify to a job, instead of doing so unconditionally.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit 7b11443ebcb00241b15517f4348d154db8056d1c
Author: Klaus Aehlig <[email protected]>
Date: Wed Jul 2 12:55:54 2014 +0200
Make max_tracked_jobs queryable
As we have introduced a new cluster parameter, it should
be also visible when querying about the cluster configuration.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit f08b23e77bae475a5800883d56c80b8309287c75
Author: Klaus Aehlig <[email protected]>
Date: Wed Jul 2 12:39:15 2014 +0200
Add a command-line parameter for max_tracked_jobs
...so that this opcode parameter can become available for 'gnt-cluster
modify'.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit 3acd363ac28faf10d9bc16c97c65110710ca5c73
Author: Klaus Aehlig <[email protected]>
Date: Tue Jul 1 17:56:08 2014 +0200
Extend OpClusterSetParams by pMaxTrackedJobs
...so that max_tracked_jobs becomes a run-time tunable.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit 34444b403031bd0da972babd1c887b4ce6f4abe6
Author: Klaus Aehlig <[email protected]>
Date: Tue Jul 1 17:53:40 2014 +0200
Add an OpCode parameter for max_tracked_jobs
...so that we can have an OpCode setting this parameter.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
commit 056405115be27cf83fe7c44d0ad7106d4306c6ed
Author: Klaus Aehlig <[email protected]>
Date: Tue Jul 1 16:36:24 2014 +0200
Add parameter max_tracked_jobs to the cluster configuration
This cluster-wide parameter will determine how many running
jobs luxid is willing to trace via inotify for scheduling
purpose. If more jobs are running, the progress of the excess
jobs will only be noticed at the periodic polling.
Signed-off-by: Klaus Aehlig <[email protected]>
Reviewed-by: Petr Pudlak <[email protected]>
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
