NexentaStor 4.0 HA Plugin Release note
The following enhancements have been made to the RSF-1 HA Plugin package for NexentaStor 4.0
Fixed stmfha core dump bug
There was a bug that caused stmfha to core dump. This happened after it had finished the functional part of the program, so didn’t affect anything.
Fix bug with disk HB suspend
Bug where old HBs were cached
MHDC problem disk detection
Some disks have been found to behave incorrectly when reserved in a multipathed system (return ‘reservation conflict’ when given a TUR command. MHDC has been updated to test for this and not set failfast on those disks (avoid kernel panic)
To further avoid the problem described above, a blacklist mechanism has been implemented, which means failfast is never set on blacklisted disk types.
MHDC improved logging !
Option to disable fping test
Before starting a service, RSF-1 uses the fping binary to check if the floating VIP is available elsewhere on the network. If it is, then the service moves to ‘broken_unsafe’ rather than starting.
This change gives the option to disable this functionality through a database update.
Now includes IPV6 support
‘Stretch cluster’ extensions
RSF-1 now has extra functionality for clusters where the nodes are separated geographically. This extra functionality allows for external relays to be used to give extra intelligence to network heartbeats. This can be enabled with an extended license.
config_dist will now retry once if it gets a connection refused
Log when client calls are blocked
rsfmon blocks any client requests (rsfcli, etc.) if the limit of 50 per poll period is reached. It now adds a message to the log file when this happens.
rsfmon call handler updated to allow for faster call rates
- START_MONITOR expressions added
MACHINE names can now be included in monitor expressions !
Added the following utilities:
- MPxIO status program from Pascal Gienger • arcstat.pl
- ZFS ARC status program from Richard Elling • zilstat.ksh
- ZFS ZIL status program from Richard Elling • cpool
- Wrapper around mhdc to reserve/release/interrogate all disks in a service