There are following error messages in ASM alert log file:
Errors in file /u01/app/grid/diag/asm/+asm/+ASM1/trace/+ASM1_m000_39281.trc (incident=198012): ORA-00603: ORACLE server session terminated by fatal error ORA-27504: IPC error creating OSD context ORA-27300: OS system dependent operation:sendmsg failed with status: 105 ORA-27301: OS failure message: No buffer space available ORA-27302: failure occurred at: sskgxpsnd2 opidrv aborting process M000 ospid (39281) as a result of ORA-603 2019-03-12T01:00:18.919871+11:00 Process m000 died, see its trace file
+ASM1_m000_39281.trc:
...
..
.
SKGXP:[7f620a621930.0]{0}: SKGXPVFYNET: Socket self-test could not verify successful transmission of 32768 bytes (mtype 61).
SKGXP:[7f620a621930.1]{0}: The network is required to support UDP protocol sends of this size. Socket is bound to 169.254.174.228.
SKGXP:[7f620a621930.2]{0}: phase 'send', 0 tries, 100 loops, 13590 ms (last)
struct ksxpp * ksxppg_ [0x7f620a68a770, 0x7f6204ff1350) = 0x7f6204ff1348
...
..
.
Workaround
1) Shrink database instance SGA size to give more memory back to OS.
After OS gets more available memory, the issue is gone.
Physical memory utilisation( %):
OR
2) Oracle has reported this issue in Doc ID 2041723.1. by changing the MTU of the loopback interface, and changing the value of kernel parameter min_free_kbytes.
a) Lower MTU to 16436 by adding following to /etc/sysconfig/network-scripts/ifcfg-lo.
MTU=16436
Then restart the network service.
# systemctl restart network.service
b) Increase the value of vm.min_free_kbytes to 0.4%of the total physical memory of the server. This can be done by adding following to /etc/sysctl.conf.
vm.min_free_kbytes = 2097152
and then run below command to make it effective:
#sysctl -p
Make DBA Life Easy 