My db failed to load and died with reason invalid_header_extension, how to relax again?

[codinggirl]

Hi, every one.

One of my db could not be accessed. The Fauxton interface shows This database failed to load. and the systemd logs shows:

5月 26 09:00:15 carry couchdb[1945]: [notice] 2022-05-26T01:00:15.992540Z couchdb@127.0.0.1 <0.17816.1> e81d8d20b0 127.0.0.1:5984 127.0.0.1 admin GET /csdnblog 500 ok 8778
5月 26 09:00:15 carry couchdb[1945]: [error] 2022-05-26T01:00:15.992216Z couchdb@127.0.0.1 <0.26577.1> -------- CRASH REPORT Process  (<0.26577.1>) with 2 neighbors crashed with reason: {invalid_header_extension,{{comp_header,{db_header,8,31494551,0,nil,{3165677038,25730618,1269653504},nil,n>
5月 26 09:00:15 carry couchdb[1945]: [info] 2022-05-26T01:00:15.992156Z couchdb@127.0.0.1 <0.266.0> -------- db shards/00000000-7fffffff/csdnblog.1652814365 died with reason {{invalid_header_extension,{{comp_header,{db_header,8,31494551,0,nil,{3165677038,25730618,1269653504},nil,nil,nil,3801>
5月 26 09:00:15 carry couchdb[1945]: [error] 2022-05-26T01:00:15.992080Z couchdb@127.0.0.1 <0.26577.1> -------- CRASH REPORT Process  (<0.26577.1>) with 2 neighbors crashed with reason: {invalid_header_extension,{{comp_header,{db_header,8,31494551,0,nil,{3165677038,25730618,1269653504},nil,n>

Access via curl with _all_docs endpoint shows:

$ curl -H "content-type: application/json"  -X GET http://admin:admin@127.0.0.1:5984/csdnblog/_all_docs
{"error":"badmatch","reason":"{{invalid_header_extension,\n     {{comp_header,\n          {db_header,8,31494551,0,nil,\n              {3165677038,25730618,1269653504},\n              nil,nil,nil,3801903443,1000,\n              <<\"d64ed202c3e8215523b7fcde7f4f851e\">>,\n              [{'couchdb@127.0.0.1',0}],\n              27999493,1000,3801903512},\n          [{root,{[5217494319],nil}},{num_kvs,25730618},{num_bb,7815}]},\n      {comp_header,\n          {db_header,8,31494551,0,nil,\n              {3165677038,25730618,1269653504},\n              nil,nil,nil,3801903443,1000,\n              <<\"d64ed202c3e8215523b7fcde7f4f851e\">>,\n              [{'couchdb@127.0.0.1',0}],\n              27999493,1000,3801903512},\n          [{root,{[5217494319],nil}},{num_kvs,25730618},{num_bb,7815}],\n          0,nil,nil,nil,nil,nil,nil,1000,undefined,undefined,undefined,1000,\n          undefined}}},\n [{couch_bt_engine_header,upgrade_tuple,1,\n      [{file,\"src/couch_bt_engine_header.erl\"},{line,233}]},\n  {lists,foldl,3,[{file,\"lists.erl\"},{line,1267}]},\n  {couch_bt_engine,init_state,4,[{file,\"src/couch_bt_engine.erl\"},{line,875}]},\n  {couch_bt_engine,init,2,[{file,\"src/couch_bt_engine.erl\"},{line,167}]},\n  {couch_db_engine,init,3,[{file,\"src/couch_db_engine.erl\"},{line,775}]},\n  {couch_db_updater,init,1,[{file,\"src/couch_db_updater.erl\"},{line,43}]},\n  {proc_lib,init_p_do_apply,3,[{file,\"proc_lib.erl\"},{line,226}]}]}","ref":3459336757}

This is a db contains about 50M docs, what's the best options to make the db accessible again? Is there a simple tool to validate and fix this?

[nickva]

@codinggirl try to first back up the database files. Then see if there are files with a .compact.meta and .compact.data extensions and delete those. Then try to restart the server.

Is there a chance this happened during upgrade (the db was created with an older CouchDB then opened with a newer CouchDB?).