Commits

Author Commit Message Labels Comments Date
Enis Afgan
Can't create a database role until the database exists...
Enis Afgan
Provisions to support FTP data upload for Galaxy
Enis Afgan
Some terminology updates dealing with instance sharing
Enis Afgan
Added ability for users to share a snapshot of their cluster. This also enables users to take snapshots of the cluster for themselves. Several UI/logging improvements.
Enis Afgan
As part of saving the cluster configuration, save an additional file in cluster's bucket whose name is the name of the cluster
Enis Afgan
Links for service control on admin panel are all functional now
Enis Afgan
Tune the volume detaching process to include a second attempt at detaching; rebooting master instance no longer detaches volumes but simply shuts down services and reboots the instance - much faster (use 'Terminate cluster' with manual instance reboot if need complete reboot)
Dannon Baker
Fixed admin panel start/stop galaxy. Added a single restart_galaxy option.
Enis Afgan
Added support for nginx configuration to allow partial content (streaming)
Enis Afgan
Rewrote composiiton of the file used for modifying SGE's allhosts group. The data in the file is now based on current instances known to master rather than parsing SGE's qconf output. This fixes occasional bug that would result in failure of instance removal from SGE's execution host list.
Enis Afgan
Added a delay before checking for a newly started instance; also, if an instance that is otherwise unknown to the master sends a message, master will add it to its instance list
Enis Afgan
Added volume and file system introspection code to ensure consistency between references for a volume and a file system (applies to a live instance and cluster config)
Enis Afgan
Fixed cluster deletion option in case the master instance was being terminated automatically at cluster shutdown
Enis Afgan
Get EC2 connection for a specific region (i.e., non-default and the same as the instance is running in). Also improved some of the logging explicitness.
Enis Afgan
An attempt at making the code for attaching volumes more robust to failing devices
Enis Afgan
Replaced some of the text intended for testing only
Enis Afgan
Bug fix when starting a new cluster; also cleaned up logging in boot script
Enis Afgan
Tweaks to messages during volume detaching (if there are problems detaching)
Enis Afgan
Added explicit update of snapshot id a volume was created from
Enis Afgan
Getting rid of erroneous print statements left over from debugging.
Enis Afgan
Adjust autoscaling to keep max number of instances if more than the max number are running when autoscaling is turned on. Also, don't try to unmount filesystems that are not in state RUNNING.
Enis Afgan
Make sure we have a complete Galaxy cluster configuration at start and not to start a partial set of services.
Dannon Baker
Merge.
Dannon Baker
Snapshot progress takes into account "Configuring" stage and presents it in the dialog. Also, much snappier.
Enis Afgan
Fix for when the CloudMan update What's new link
Dannon Baker
Snapshot status work.
Dannon Baker
Cluster Reboot through UI w/reloading dialog, linked in with CM Update.
Enis Afgan
Compose 'What's new' link for updating CloudMan to point to the exact set of changesets on bitbucket
Enis Afgan
A more complete 'cluster delete' approach where any volumes tagged with cluster name are deleted upon cluster deletion
Enis Afgan
Had to postpone instance reboot after master restart because NFS was not ready by the time a worker would come up.
  1. Prev
  2. Next