A quick and simple way to check the healthy of your IB Switches in Exadata is this one below:
1-Check using this command below:
ibqueryerrors.pl -rR -s PortRcvSwitchRelayErrors,PortXmitDiscards,PortXmitWait,VL15Dropped
2-If you got something line that:
GUID 0x10e040687aa0a0 port ALL: [LinkErrorRecoveryCounter == 152] [PortRcvErrors == 5]
GUID 0x10e040687aa0a0 port 1: [LinkErrorRecoveryCounter == 12]
and this value is not increasing a lot, you can clear these errors and continue with your life without any problem but if this number are increasing quickly or are too high (30k, 65k or more) it is better to open an SR to check the IB hardware.
3-Clear the count so when you run some tool like RACchk or the new ORAchk you will not be warned about these little errors.
[root@ex01dbadm01 ORAchk]# ibclearcounters
## Summary: 25 nodes cleared 0 errors
[root@ex01dbadm01 ORAchk]#
Best Regards,
Paulo Portugal
1-Check using this command below:
ibqueryerrors.pl -rR -s PortRcvSwitchRelayErrors,PortXmitDiscards,PortXmitWait,VL15Dropped
2-If you got something line that:
GUID 0x10e040687aa0a0 port ALL: [LinkErrorRecoveryCounter == 152] [PortRcvErrors == 5]
GUID 0x10e040687aa0a0 port 1: [LinkErrorRecoveryCounter == 12]
and this value is not increasing a lot, you can clear these errors and continue with your life without any problem but if this number are increasing quickly or are too high (30k, 65k or more) it is better to open an SR to check the IB hardware.
3-Clear the count so when you run some tool like RACchk or the new ORAchk you will not be warned about these little errors.
[root@ex01dbadm01 ORAchk]# ibclearcounters
## Summary: 25 nodes cleared 0 errors
[root@ex01dbadm01 ORAchk]#
Best Regards,
Paulo Portugal
