Issue #2283: Errors after unexpected server shutdown

After an unexpected server shutdown and restart of our application we have these messages in the log (complete log attached):
…

[2018-03-26 14:39:30 #1096 store] 
Request Value: 10:'b61e33d4-8d57-478d-9176-6aa0e2031470'


[2018-03-26 14:39:30 #1097 store] 
*** Attempt to remove a non existing index entry:

[2018-03-26 14:39:30 #1098 store] 
BTree ID: -170


[2018-03-26 14:39:30 #1099 store] 
BTree Name: com.agile.hummingbird.ObjectNode[modified]

[2018-03-26 14:39:30 #1100 store] 
BTree Page: 18071876

[2018-03-26 14:39:30 #1101 store] 
Request code: 4

[2018-03-26 14:39:30 #1102 store] 
Request Key: 2018-03-26 13:26:27

[2018-03-26 14:39:30 #1103 store] 
Request Key Code: -30


[2018-03-26 14:39:30 #1104 store] 
Request Real Key: 2018-03-26 13:26:27

[2018-03-26 14:39:30 #1105 store] 
Request Real Key Code: -30


[2018-03-26 14:39:30 #1106 store] 
Request Small Key: 2018-03-26 13:26:27

[2018-03-26 14:39:30 #1107 store] 
Request Small Key Code: -30


[2018-03-26 14:39:30 #1108 store] 
Request Value: 10:'19f9ec33-84a9-4957-9379-79b27c688bb6'

...

again there is no possibility to enter a ticket, on pressing save we have this message:

The log indeed shows that the database was recovered from an unexpected shutdown. Unfortunately it seems that one of the indexes (ObjectNode.modified) became non synchronized with the data or corrupted. This shouldn't happen of course as recovery from failure should protect data and indexes, so we will have to investigate this further.

As a quick solution, you may try to remove the index, i.e. access the database with a class (in the classpath)  in which that index is not defined, and then again redefine the index by accessing the database with a class (in the classpath) in which that index is defined.

Alternatively you should be able to repair the entire database by running the Doctor.

We will also check why you got the error in submitting the ticket.

Further discussion of this issue is in issue #2284.