Is it OK if you just restart with the 2.2.2 master?
Adam Sjøgren 于2024年3月25日周一 20:14写道:
>
> Hi,
>
> I am trying to do a rolling upgrade of an HBase 2.2.2 cluster to
> 2.5.4.
>
> When I try switching over to the new master, it processes the WAL
> files and then ends up saying:
>
> [... many lines elided ...]
>
> 2024-03-25T12:44:29,374 ERROR [master/s4:16000:becomeActiveMaster]
> region.RegionProcedureStore: Corrupted procedure pid=221160,
> ppid=219241, state=RUNNABLE; OpenRegionProcedure
> 47b69716553a2e2587f22e0e7c3d268a, server=node13,16020,1640161141416
>
> 2024-03-25T12:44:29,374 ERROR [master/s4:16000:becomeActiveMaster]
> region.RegionProcedureStore: Corrupted procedure pid=221161,
> ppid=219258, state=RUNNABLE; OpenRegionProcedure
> 1071a48e8ab7730a160cf55391d1670a, server=node23,16020,1639980983954
>
> 2024-03-25T12:44:29,374 ERROR [master/s4:16000:becomeActiveMaster]
> region.RegionProcedureStore: Corrupted procedure pid=4215645,
> ppid=4215619, state=SUCCESS, bypass=true; OpenRegionProcedure
> 17c468cbf507e548b1c34c74b14b4c40, server=node09,16020,1700398035068
>
> 2024-03-25T12:44:29,375 ERROR
> [master/s4:16000:becomeActiveMaster] master.HMaster: Failed to
> become active master
> java.io.IOException: There are 2336 corrupted procedures when
> migrating from the old WAL based store to the new region based
> store, please fix them before upgrading again.
>
> and then it quits.
>
> How do I fix these 2336 corrupted procedures?
>
> Best regards,
>
> Adam
>
> --
> "Why does the sky turn red as the sun sets?" Adam Sjøgren
> "That's all the oxygen in the atmosphere catching a...@koldfront.dk
> fire."
>