hpssi.ppt

1

• HPs OpenSSI Linux Cluster Project
• www.openssi.org
• Bruce J. Walker
• HP FellowOffice of Strategy and Technology

2
Overview of HP OpenSSI Cluster

• Single HA root filesystem
• Consistent OS kernel on each node
• Join cluster early in boot
• Strong Membership
• Single view of filesystems, devices, processes,
  ipc objects
• Single management domain
• Load balancing of connections and processes
• Dynamic service provisioning

3
Goals

• Become the Industry Standard in Linux Clustering
• Scale up and Scale down
• Modular technology that applies to all forms of
  clustering
• - single technology base for HPC and Commercial

4
Clustering Goals

• To various degrees, all of
• High Availability
• Scalability
• Manageability
• Usability

5
HP OpenSSI Linux Clusters
Ideal/Perfect Cluster in all dimensions
SMP
SMP
Typical HA Cluster
HP SSI Linux Clusters
log scale
HUGE
ReallyBIG
6
HP OpenSSI Linux ClustersApproach

• 2 or more service nodes optional compute
  nodes
• Computing can be done on service nodes
• Service nodes provide availability and 2 forms of
  load balancing
• Compute nodes for larger scale
• Vproc for clusterwide process space, process
  launch and process movement
• Lustre for scale up filesystem story (including
  root)
• CFS or GFS for scale down
• Diskless node option
• No daemons need on compute nodes for job launch
  or stdio
• Integration with other open source components
• Lustre, LVS, PBS, Maui, Ganglia, SuperMon, SLURM,
  

7
Availability

• No Single (or even multiple) Point(s) of Failure
• Automatic Failover/restart of services in the
  event of hardware or software failure
• Filesystem failover integrated and automatic
• Application Availability is simpler in an SSI
  Cluster environment statefull restart easily
  done
• Application transparent checkpoint/restart
• Online software upgrade (ongoing)
• Architected to avoid scheduled downtime
• Node eviction via transparent process migration

8
HP OpenSSI Linux Clusters -Manageability

• Single Installation single root filesystem
• Joining the cluster is automatic as part of
  booting and doesnt have to managed
• Trivial online addition of new nodes
• Use standard single node tools (SSI Admin)
• Visibility of all resources of all nodes from any
  node
• Applications, utilities, programmers, users and
  administrators often neednt be aware of the SSI
  Cluster
• Simpler HA (high availability) management

9
HP OpenSSI Linux ClusterEase of Use

• Can run anything anywhere with no setup
• Can see everything from any node
• service failover/restart is trivial
• automatic or manual load balancing
• powerful environment for application service
  provisioning, monitoring and re-arranging as
  needed

10
Scalability

• Two aspects overhead and load balancing

11
Scalability - Overhead

• No daemons needed per node
• (single HA clusterwide init)
• Node monitoring adjustable and could co-ordinate
  noise
• Minimal cross-node co-ordination
• Efficient kernel-to-kernel rexec() for process
  launch
• Parallel multicast boot

12
Scalability Load Balancing

• Load balancing of connections and processes
• SSI makes distributing function very easy
• SSI allows any process on any processor
• All resources transparently visible from all
  nodes
• filesystems, IPC, processes, devices,
  networking
• OS version in local memory on each node
• Migrated processes use local resources and not
  home-node resources

13
Component Contributions to HP OpenSSI Cluster
Project
Lustre
Appl. Avail.
CLMS
GFS
Mosix
Beowulf
Vproc
DLM
LVS
OCFS
IPC
DRBD
CFS
HP OpenSSI Cluster Project
EVMS
Load Leveling



HP contributed
Open source and integrated
To be integrated
14
SSI Cluster Architecture/ Components
18. Timesync
14. Init booting run levels
13. Packaging and Install
15. Sysadmin
16. Appl Availability HA daemons
17. Application Service Provisioning
19. MPI, etc.
Kernel Interface
3. Filesystem
6. IPC
5. Process Loadleveling
1. Membership
CFS
GFS
Physical filesystems
7. Networking/ LVS
4. Process Mgmt VPROC
Lustre
9. Devices/ shared storage devfs
8. DLM
10. Kernel data replication service
11. CLVM/ EVMS (TBD)
2. Internode Communication/ HA interconnect
12. DRBD
15
Vproc Features

• Clusterwide unique pids (decentralized)
• Peer-to-peer no master node
• process and process group tracking under
  arbitrary failure and recovery
• no polling
• reliable signal delivery under arbitrary failure
• process always executes system calls locally
• no do-do at home node never more than 1 task
  struct per process
• for HA and performance, processes can completely
  move
• therefore can service a node without application
  interruption
• process always only has 1 process id
• transparent process migration
• clusterwide /proc,
• clusterwide job control
• single init
• Unmodified ps shows all processes on all nodes
• transparent clusterwide debugging (ptrace or
  /proc)
• integrated with load leveling (manual and
  automatic)
• exec time and migration based automatic load
  leveling
• fastnode command and option on rexec, rfork,
  migrate

16
HP SSI Linux Clusters - Status

• Version 1.0 just released
• Binary, Source and CVS options
• Functionally complete RH9 and RHel3
• Debian release also available
• IA-32 and Itanium Platforms
• Runs HPTC apps as well as Oracle RAC
• Available at OpenSSI.org

17
Ongoing Enhancements

• 2.6-based version
• Improved scalability (two node classes)
• Integrate power/console management
• SMP Clusters (multiple nodes within a large SMP)
• Vproc without single root (flexibility)
• Node eviction

18
HP OpenSSI Linux Clusters - Conclusions

• HP has recognized that Linux clusters are
  important part of the future.
• HP has recognized that combining scalability with
  availability and manageability/ease-of-use is
  key to clustering
• HP is leveraging its merger with Compaq
  (Tandem/Digital) to bring the very best of
  clustering to a Linux base