MongoDB MMS 副本集無法重啟

我有一個部署了 3 個實例的 MMS 副本集。它工作正常，直到今天早上我將其關閉以進行一些更改並重新啟動它。從日誌看，感覺副本集有不干淨的關閉。主伺服器上的 mongod 伺服器現在無法啟動。

這是完整的日誌

2015-08-20T07:08:41.389+0000 W -        [initandlisten] Detected unclean shutdown - /data/XXXXXXXXXXXX/mongod.lock is not empty.
2015-08-20T07:08:41.406+0000 I JOURNAL  [initandlisten] journal dir=/data/XXXXXXXXXXXXX/journal
2015-08-20T07:08:41.406+0000 I JOURNAL  [initandlisten] recover begin
2015-08-20T07:08:41.406+0000 I JOURNAL  [initandlisten] info no lsn file in journal/ directory
2015-08-20T07:08:41.406+0000 I JOURNAL  [initandlisten] recover lsn: 0
2015-08-20T07:08:41.406+0000 I JOURNAL  [initandlisten] recover /data/XXXXXXXXXXXX/journal/j._0
2015-08-20T07:08:41.407+0000 I JOURNAL  [initandlisten] recover cleaning up
2015-08-20T07:08:41.407+0000 I JOURNAL  [initandlisten] removeJournalFiles
2015-08-20T07:08:41.641+0000 I JOURNAL  [initandlisten] recover done
2015-08-20T07:08:41.641+0000 I JOURNAL  [initandlisten] preallocating a journal file /data/XXXXXXXXXXX/journal/prealloc.0
2015-08-20T07:08:44.074+0000 I -        [initandlisten]   File Preallocator Progress: 744488960/1073741824 69%
2015-08-20T07:08:47.176+0000 I -        [initandlisten]   File Preallocator Progress: 901775360/1073741824 83%
2015-08-20T07:08:50.274+0000 I -        [initandlisten]   File Preallocator Progress: 1027604480/1073741824 95%
2015-08-20T07:09:09.057+0000 I JOURNAL  [durability] Durability thread started
2015-08-20T07:09:09.057+0000 I JOURNAL  [journal writer] Journal writer thread started
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] MongoDB starting : pid=25507 port=27000 dbpath=/data/XXXXXXXXXXXXX 64-bit host=CH$
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] db version v3.0.2
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] git version: 6201872043ecbbc0a4cc169b5482dcf385fc464f
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] OpenSSL version: OpenSSL 1.0.1f 6 Jan 2014
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] build info: Linux ip-10-229-1-2 3.13.0-24-generic #46-Ubuntu SMP Thu Apr 10 19:11:08 U$
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] allocator: tcmalloc
2015-08-20T07:09:09.060+0000 I CONTROL  [initandlisten] options: { config: "/data/XXXXXXXXXXX/automation-mongod.conf", net: { port: 270$
2015-08-20T07:09:09.074+0000 I -        [initandlisten] Invariant failure _name == nsToDatabaseSubstring( ns ) src/mongo/db/catalog/database.c$
2015-08-20T07:09:09.091+0000 I CONTROL  [initandlisten]
0xf4f859 0xef0031 0xed4b52 0x91e106 0x91e18f 0x920033 0x922cb0 0x808701 0x7d4ba4 0x7f0503489ec5 0x805d17
----- BEGIN BACKTRACE -----
{"backtrace":[{"b":"400000","o":"B4F859"},{"b":"400000","o":"AF0031"},{"b":"400000","o":"AD4B52"},{"b":"400000","o":"51E106"},{"b":"400000","o$
mongod(_ZN5mongo15printStackTraceERSo+0x29) [0xf4f859]
mongod(_ZN5mongo10logContextEPKc+0xE1) [0xef0031]
mongod(_ZN5mongo15invariantFailedEPKcS1_j+0xB2) [0xed4b52]
mongod(_ZNK5mongo8Database13getCollectionERKNS_10StringDataE+0x36) [0x91e106]
mongod(_ZN5mongo8Database30_getOrCreateCollectionInstanceEPNS_16OperationContextERKNS_10StringDataE+0x1F) [0x91e18f]
mongod(_ZN5mongo8DatabaseC1EPNS_16OperationContextERKNS_10StringDataEPNS_20DatabaseCatalogEntryE+0x1E3) [0x920033]
mongod(_ZN5mongo14DatabaseHolder6openDbEPNS_16OperationContextERKNS_10StringDataEPb+0x150) [0x922cb0]
mongod(_ZN5mongo14DatabaseHolder6openDbEPNS_16OperationContextERKNS_10StringDataEPb+0x150) [0x922cb0]
mongod(_ZN5mongo13initAndListenEi+0xC01) [0x808701]
mongod(main+0x134) [0x7d4ba4]
libc.so.6(__libc_start_main+0xF5) [0x7f0503489ec5]
mongod(+0x405D17) [0x805d17]
-----  END BACKTRACE  -----
2015-08-20T07:09:09.091+0000 I -        [initandlisten]

***aborting after invariant() failure

知道如何解決嗎？自過去 4 小時以來，我一直在嘗試修復它，但似乎沒有任何效果。

因此，事實證明問題出在.ns其中一個數據庫的文件中。我不得不刪除該文件並重新啟動伺服器。伺服器啟動成功，但是我刪除的 .ns 數據庫失去了。

對於 mongoDB 中的新手來說，.ns 文件是 mongo 在其 data ( /data) 目錄中為每個數據庫創建的命名空間文件。

引用自：https://dba.stackexchange.com/questions/111650