This page tracks projects/ideas/features which are targeted towards GSoC 2017. We want to consolidate all the projects so that it is easier for people to choose.
Note to potential candidates: If you like any of the projects, shoot a mail to dev@cloudstack.apache.org with the JIRA ID mentioned in it.
JIRA ID | Title | Short Description | Potential Mentors | Candidates Interested |
---|---|---|---|---|
CLOUDSTACK-9804 | Add Cinder as a storage driver to Cloudstack | The idea is to add cinder support for managed storage. Cinder is *The* platform for doing Software defined storage. Every vendor has plugged in their device to Cinder. We can run Cinder standalone [1] . This would mean that once we integrate Cinder, we get all the other storage vendor support for free, NetApp, HP ... the list is huge [2] [1] https://gorka.eguileor.com/standalone-cinder/ [2] https://wiki.openstack.org/wiki/CinderSupportMatrix | Syed Ahmed | Shanika Ediriweera |
CLOUDSTACK-9778 | Replace custom console with NoVNC console | The idea is to relplace our old and crummy console with a NoVNC console. There are a lot of advantages of using NoVNC
| Syed Ahmed | Shanika Ediriweera |
Decouple Cloudstack UI | Just like cloudmonkey, decouple Cloudstack UI to a separate project. It should be able to talk to any cloudstack endpoint | Rajani Karuturi | Shanika Ediriweera | |
Use Ceph RBD storage pool for writing HA heartbeats | Just like NFS we should write a heartbeat for each Instance to RADOS. Each hosts could write a simple object like: <pool>/<host ip> They simply write the timestamp to the object encoded in JSON. Other hosts can read that object and see if the host wrote the timestamp lately. If it did it means that it is still up and running and Fencing is not needed or required. | Wido den Hollander | Sachin Patil | |
commit to commit db upgrades and db version control | Cloudstack currently uses a homegrown tool to do the database upgrades. To achieve this, we can use existing and well tested tools like liquibase, flywaydb etc. or improve on the existing one. Related discussions on dev lists | Rajani Karuturi | ||
Add TRIM/Discard support to Qemu | Thinly provisioned volumes on storage devices continue to grow because the storage device has no idea of which blocks are in use. For SSDs the TRIM/Discard feature was invented to give free/unused blocks back to the flash device, but this can also be used for Qemu. Ceph's RBD for example supports trimming so that volumes can shrink again when blocks are no longer in use. This is supported since Qemu 1.5, but since 1.6 it also works for QCOW2 images. It however requires the new virtio-scsi to work optimal, so it requires some changes. For more information see: | Wido den Hollander | Sachin Patil |