HA/FST
HA/FST is a high availability clustering tool for up to four Solaris nodes. It supports any application that can be managed from the command line. It comes with support for NIC failover, Apache, Oracle, NFS, Samba, Solaris Volume Manager, Veritas Volume Manager, and application fallback and failto.
| Tags | Clustering/Distributed Networks Systems Administration |
|---|---|
| Licenses | GPL |
| Operating Systems | POSIX Solaris |
Recent releases
- 12 Jan 2006 20:01
Changes: The requirement for a license to use extended features was removed, and the source code was released under the GNU GPL. If virtual host takeover occurred on the last node in the cluster, the bringup is no longer liable to be attempted on any random node. Using a reservation disk no longer results in bringing the cluster down due to remaining nodes' inability to lock a disk until SCSI timeout occurs. Label corrections in fsthamon and minor documentation corrections were made.
- 11 Jan 2006 19:23
Changes: A bug where a syntax error parsing fstmodapp's -t option prevented the proper creation of application agents was fixed.
- 09 Jan 2006 10:26
Changes: The requirement for a license to use extended features was removed, and the source code was released under the GNU GPL. If virtual host takeover occurred on the last node in the cluster, the bringup is no longer liable to be attempted on any random node. Using a reservation disk no longer results in bringing the cluster down due to remaining nodes' inability to lock a disk until SCSI timeout occurs. Label corrections in fsthamon and minor documentation corrections were made.
- 22 Oct 2005 07:55
Changes: Process management has been reworked, resulting in a major performance improvement. The heartbeat subsystem has been reworked and now uses UDP packets instead of ICMP pings, and can be configured for load. Refreshing in fsthatool has been automated, and the GUI has been updated. The ability to choose which node to failover to has been included, along with an option to monitor applications from other clusterd nodes and receive notifications for cluster failures. An optional ethernet address has been added to nettab.
- 22 Oct 2005 01:25
Changes: Process management was reworked, resulting in a major performance improvemnt. The heart beat subsystem was also reworked and now uses UDP packets instead of ICMP pings and can be configured for load. Refresh in fsthatool was automated and the GUI was updated. The ability to choose which node to failover to was included along with an option to monitor applications from other clusterd nodes, and receive notifications for cluster failures. An optional ethernet address was added to nettab.
