Restore RT Index

I am in a bit of a strange position. After one of the reboots, my RT index fails to be recognized by manticore. It is as if it wasn’t there at all.
I can see that the data files are still there, apparently intact.
Index was created via SQL, there is no index configuration for it in config file, only searchd section there.
Any reason why this would happen? Can I recover from it?

could you check manticore.json file at the data directory?

could you check searchd.log for any error or warning messages related to index?

manticore.json looks like this:
{
“clusters”: {
},
“indexes”: {
}
}

searchd.log has a lot of errors like this:
[Sun Jul 23 12:59:08.141 2023] [1303] starting daemon version ‘3.6.0 96d61d8bf@210504 release (columnar 1.0.0 768b007@210503)’ …
[Sun Jul 23 12:59:08.141 2023] [1303] listening on all interfaces for sphinx and http(s), port=9412
[Sun Jul 23 12:59:08.141 2023] [1303] listening on all interfaces for mysql, port=9406
[Sun Jul 23 13:01:31.854 2023] [1310] FATAL: disk chunk /path/index.10945: prealloc failed: failed to mmap file ‘/path/index.10945.spa’: Cannot allocate memory (length=494260)
[Sun Jul 23 13:01:31.854 2023] [1366] WARNING: GlobalCrashQueryGetRef: thread-local info is not set! Use ad-hoc
[Sun Jul 23 13:01:31.861 2023] [1301] watchdog: got error 4, Interrupted system call
[Sun Jul 23 13:01:35.252 2023] [1301] watchdog: main process 1303 killed dirtily with signal 11, will be restarted
[Sun Jul 23 13:01:35.252 2023] [1301] watchdog: got error 10, No child processes
[Sun Jul 23 13:01:35.252 2023] [1301] watchdog: main process 1948 forked ok

ending with

[Sun Jul 23 15:00:54.726 2023] [31762] WARNING: index ‘index’: removed from JSON config
[Sun Jul 23 15:00:54.726 2023] [31762] WARNING: index ‘index’: prealloc: failed to lock /path/index.lock: Resource temporarily unavailable - NOT SERVING
[Sun Jul 23 15:00:54.751 2023] [31761] accepting connections
[Sun Jul 23 15:00:54.760 2023] [31772] prereading 0 indexes
[Sun Jul 23 15:00:54.760 2023] [31772] prereaded 0 indexes in 0.000 sec

seems you have no free memory to load the index and daemon got killed by OS

You could add index by hand when daemon shutdown into maticore.json like

	"indexes":	{
		"t11":	{
			"type":	"rt",
			"path":	"t11"
		},
		"t12":	{
			"type":	"rt",
			"path":	"t12"
		},
...

However OS still could kill the daemon.