Hi folks,
I encountered an exception which is index recover failed due to translog EOFexception, I did some research and it seems to be corrupted translog.
I tried to delete the translog directory(and all files contained) of the problematic index, then restart the node. The exception popuped as below:
[2018-05-03T10:03:27,282][WARN ][o.e.c.a.s.ShardStateAction] [0fcbafcc-e892-468a-820c-a641a49107d1] [data_20180504][0] received shard failed for shard id [[data_20180504][0]], allocation id [t0r1USN7RlmfI2tcqrvyVA], primary term [0], message [failed recovery], failure [RecoveryFailedException[[data_20180504][0]: Recovery failed on {0fcbafcc-e892-468a-820c-a641a49107d1}{LxzCvECiTZOzzTfFL7NYKg}{VGPw5nugQN6COcRlruao5A}{172.17.11.11}{172.17.11.11:9300}]; nested: IndexShardRecoveryException[failed to recover from gateway]; nested: EngineCreationFailureException[failed to create engine]; nested: NoSuchFileException[/ES/data/xyz/nodes/0/indices/TMCdiDCpR62Z5HzwoEB-HA/0/translog/translog.ckp]; ]
My question:
-
Delete the translog directory is a right approach to deal with this problem at first? I don't care lost those docs in translog.
-
For the current situation, how to recover?
Thanks for you sharing.