Hello,
I would like to know if anyone has faced this problem. Whenever there is a duplicate IP address, HACMP goes down infact HACMP ( PowerHA ) takes the whole system down.
Does anyone know how to solve this problem ?
Just avoid the duplicate IP.
the problem occurs only if someone hijacks the service IP address...
is there a way i can disable in HACMP config that if there is a duplicate ip do not shutdown the system ?
---------- Post updated at 12:18 PM ---------- Previous update was at 10:49 AM ----------
on another machine, there is another strange thing going on:
# errpt|more
IDENTIFIER TIMESTAMP T C RESOURCE_NAME DESCRIPTION
1B3F9565 0103192710 P S usbhc1 SOFTWARE ERROR
AFA89905 0103192610 I O grpsvcs Group Services daemon started
923E1911 0103192610 P S topsvcs Failed to open NIM connection
97419D60 0103192610 I O topsvcs Topology Services daemon started
A6DF45AA 0103192410 I O RMCdaemon The daemon is started.
51E537B5 0103192410 P H sysplanar0 platform_dump saved to file
D221BD55 0103192410 I O perftune RESTRICTED TUNABLES MODIFIED AT REBOOT
291D64C3 0103192410 I H sysplanar0 Platform dump data
2BFA76F6 0103192110 T S SYSPROC SYSTEM SHUTDOWN BY USER
9DBCFDEE 0103192410 T O errdemon ERROR LOGGING TURNED ON
FE2DEE00 0103191110 P S SYSXAIXIF DUPLICATE IP ADDRESS DETECTED IN THE NET
BC3BE5A3 0103191110 P S SRC SOFTWARE PROGRAM ERROR
BA431EB7 0103191110 P S SRC SOFTWARE PROGRAM ERROR
12081DC6 0103191110 P S haemd SOFTWARE PROGRAM ERROR
64368504 0103191110 P O grpsvcs Connection failure between Group Service
6D19271E 0103191110 I O topsvcs Topology Services daemon stopped
AA8AB241 0103191110 T O OPERATOR OPERATOR NOTIFICATION
BC3BE5A3 0103191110 P S SRC SOFTWARE PROGRAM ERROR
AA8AB241 0103191110 T O clstrmgrES OPERATOR NOTIFICATION
F7FA22C9 0102122210 I O SYSJ2 UNABLE TO ALLOCATE SPACE IN FILE SYSTEM
F7FA22C9 0101161810 I O SYSJ2 UNABLE TO ALLOCATE SPACE IN FILE SYSTEM
4BE53A52 0101150010 P O dumpcheck Livedump filesystem almost full
HACMP resources were not moved since the system went into non-responsive state and only when one system was manually switched off resources were moved to another server. Manually switched off from the panel.
Exactly. However, preventing sbdy else using your Service Label is way out of scope for PowerHA. This task must be resolved by the server or network administrators. Hint: use the MAC address to identify the adapter that is also using the Service Label. The network guys should be able to tell you at least on what switch the other machine is connected to.
Not really as PowerHA works as designed here. In fact it is trying to protect your application's data from getting corrupted by being accessed from two different servers simultaneously. If you want to run PowerHA in an unstable environment it would be better to not use PowerHA at all instead of fiddling around with the cluster scripts.