[sheepdog-users] Sheepdog 0.6.0 released
namei.unix at gmail.com
Mon Jun 3 09:43:55 CEST 2013
On 06/03/2013 01:12 AM, MORITA Kazutaka wrote:
> Hi all,
> I'm pleased to announce the availability of the 0.6.0 release.
> The software is available from GitHub:
> Major features since 0.5.6 include:
> - Add MD support
> Sheepdog can manage multiple disks with one daemon now. The basic
> idea of MD is RAID-0 like mechanism that distributes sheep objects
> on the local disks without parity or replicating, which instead
> relies on the sheepdog's replicated storage to recover the lost
> objects on the faulty disk.
> - Add journaling support
> Yuan reworked the old journal implementation and boosted random
> write performance of sheep.
> - Move farm from sheep to collie
> The farm cluster snapshot in sheep had some drawbacks:
> 1. After a new node joins the cluster, all former snapshots are
> not available to be restored.
> 2. It is hard to backup cluster snapshot to another storage
> system for disaster-tolerance.
> 3. It is hard to initialize a new cluster by loading another
> cluster's snapshot.
> New implementation is to move "farm" from sheep to collie and save
> cluster snapshot to a local path. This retains all features from
> "farm", including:
> 1. object de-duplication
> 2. incremental store capability
> In addition, it also provides ability of:
> 1. export cluster snapshot to other storage device for backup and
> 2. deploy new cluster by restore from one snapshot of other
> - Use hash check for vdi check and recovery
> To reduce network traffic, Sheepdog uses a hash values to compare
> objects between local and remote.
> - Shepherd
> Currently, corosync doesn't have enough scalability and ZooKeeper
> is not stable under heavy network traffic. This is an initial
> implementation for yet another cluster coordinator to solve the
> Shepherd is not mature yet. Don't use it for other than testing
> - Add support for a trim command
> With this support, Sheepdog can remove the objects which are
> unnecessary for the guest virtual machines.
> The next release 0.7.0 will be in the end of August. We decided to
> use the same release cycle as QEMU. I'll announce our schedule after
> QEMU community announce theirs.
Users from 0.5.6 or old should know that object cache is now the only
implementation of cache type, so the command line is changed and no
backward compatibility for simplicity.
$ sheep -w size=200000 # enable 200G object cache
$ sheep -w size=200000,directio # enable directio mode for object cache
$ sheep -w size=20000,dir=/pato/to/object_cache
More information about the sheepdog-users