HBASE-30014: refresh_meta not working due to regionNode lock

[Kota-SH]

Jira: HBASE-30014

2026-03-24T16:27:39,332 INFO  [PEWorker-14] procedure.RefreshMetaProcedure: Executing RefreshMetaProcedure state: REFRESH_META_FINISH
2026-03-24T16:27:39,332 INFO  [PEWorker-14] util.FSTableDescriptors: Invalidating table descriptor cache.
2026-03-24T16:27:39,332 INFO  [PEWorker-14] procedure.RefreshMetaProcedure: RefreshMetaProcedure completed successfully
. All follow-up actions finished.
2026-03-24T16:27:39,333 INFO  [PEWorker-14] procedure2.ProcedureExecutor: Finished pid=4, state=SUCCESS, hasLock=false;
 RefreshMetaProcedure table=hbase:meta_replica1 in 427 msec
2026-03-24T16:27:42,670 INFO  [RpcServer.default.FPBQ.Fifo.handler=29,queue=2,port=16000] util.FSTableDescriptors: Fetc
hing table descriptors from the filesystem.

hbase:003:0>
hbase:003:0> list
TABLE
0 row(s)
Took 0.0104 seconds
=> []
hbase:004:0> refresh_meta
Refresh meta procedure submitted. Procedure ID: 4
Took 0.0141 seconds
=> 4
hbase:005:0> list
TABLE
t1
1 row(s)
Took 0.0193 seconds
=> ["t1"]
hbase:006:0>

[kgeisz]

LGTM. Although this PR doesn't add any unit tests, it does TestRefreshMetaProcedureIntegration.testRestoreTableStateForOrphanRegions()

[anmolnar]

Patch lgtm. 🚢
Could you please provide a short explanation about the root cause of the problem and why this resolution is correct?
Preferably comment on the jira ticket which will be easier to find later I think.
Is this a bugfix which is also present on the master branch or specific to our feature?
Thanks!