daicheng created KUDU-3536:
------------------------------

             Summary: Could not remove renamed recovery dir(nfs) when kudu 
restarts
                 Key: KUDU-3536
                 URL: https://issues.apache.org/jira/browse/KUDU-3536
             Project: Kudu
          Issue Type: Bug
    Affects Versions: 1.16.0
            Reporter: daicheng


Configured kudu directories to NFS on k8s , and insert some data to kudu,after 
restart kudu, the kudu tserver  fails to bootstrap with error like :
{code:java}
IO error: Could not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 One or more errors occurred {code}
while the issue didn't comes when the directory on local disk.

here some error details:
{code:java}
 Config source |        Replicas        | Current term | Config index | 
Committed?
---------------+------------------------+--------------+--------------+------------
 master        | A*  B                  |              |              | Yes
 A             | [config not available] |              |              | 
 B             | [config not available] |              |              | Tablet 
1bb9b2f91c3f48d7a97fb974112dedd6 of table 'impala::test.test_kudu' is 
unavailable: 2 replica(s) not RUNNING
  1bf087d776394884b2031385cd7e8b82 
(kudu-tserver-0.kudu-tservers.qilu-local.svc.cluster.local:7050): not running
    State:       FAILED
    Data state:  TABLET_DATA_READY
    Last status: IO error: Could not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703663028897150:
 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703663028897150:
 One or more errors occurred
  ea0e0a381c284877aa234228ed81a24f 
(kudu-tserver-1.kudu-tservers.qilu-local.svc.cluster.local:7050): not running 
[LEADER]
    State:       FAILED
    Data state:  TABLET_DATA_READY
    Last status: IO error: Could not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 One or more errors occurred{code}
{code:java}
W1227 07:43:15.222187 74 env_posix.cc:2337] Could not delete directory: IO 
error: 
/var/lib/kudu/tserver/wals/3b734a27abc74768ad6cff599b66f0f1.recovery-1703662995205917:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmW1227 07:43:15.222219 
74 env_posix.cc:2063] Error running callback with file 
/var/lib/kudu/tserver/wals/3b734a27abc74768ad6cff599b66f0f1.recovery-1703662995205917
 during walk: IO error: 
/var/lib/kudu/tserver/wals/3b734a27abc74768ad6cff599b66f0f1.recovery-1703662995205917:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmE1227 07:43:15.261075 
74 ts_tablet_manager.cc:1378] T 3b734a27abc74768ad6cff599b66f0f1 P 
ea0e0a381c284877aa234228ed81a24f: Tablet failed to bootstrap: IO error: Could 
not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/3b734a27abc74768ad6cff599b66f0f1.recovery-1703662995205917:
 
/var/lib/kudu/tserver/wals/3b734a27abc74768ad6cff599b66f0f1.recovery-1703662995205917:
 One or more errors occurredWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.261124 74 
ts_tablet_manager.cc:1356] T 3b734a27abc74768ad6cff599b66f0f1 P 
ea0e0a381c284877aa234228ed81a24f: Time spent bootstrapping tablet: real 0.213s 
user 0.070s sys 0.035sWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.261147 74 
tablet_replica.cc:323] stopping tablet replicaWed, Dec 27 2023 3:43:15 pmI1227 
07:43:15.261160 74 raft_consensus.cc:2227] T 3b734a27abc74768ad6cff599b66f0f1 P 
ea0e0a381c284877aa234228ed81a24f [term 1 FOLLOWER]: Raft consensus shutting 
down.Wed, Dec 27 2023 3:43:15 pmI1227 07:43:15.261169 74 
raft_consensus.cc:2256] T 3b734a27abc74768ad6cff599b66f0f1 P 
ea0e0a381c284877aa234228ed81a24f [term 1 FOLLOWER]: Raft consensus is shut 
down!Wed, Dec 27 2023 3:43:15 pmI1227 07:43:15.261204 74 
tablet_bootstrap.cc:492] T 1bb9b2f91c3f48d7a97fb974112dedd6 P 
ea0e0a381c284877aa234228ed81a24f: Bootstrap starting.Wed, Dec 27 2023 3:43:15 
pmI1227 07:43:15.452575 74 tablet_bootstrap.cc:492] T 
1bb9b2f91c3f48d7a97fb974112dedd6 P ea0e0a381c284877aa234228ed81a24f: Bootstrap 
replayed 1/1 log segments. Stats: ops{read=4406 overwritten=0 applied=4406 
ignored=2} inserts{seen=0 ignored=0} mutations{seen=0 ignored=0} 
orphaned_commits=0. Pending: 0 replicatesWed, Dec 27 2023 3:43:15 pmW1227 
07:43:15.469259 74 env_posix.cc:2337] Could not delete directory: IO error: 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmW1227 07:43:15.469303 
74 env_posix.cc:2063] Error running callback with file 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637
 during walk: IO error: 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmE1227 07:43:15.504146 
74 ts_tablet_manager.cc:1378] T 1bb9b2f91c3f48d7a97fb974112dedd6 P 
ea0e0a381c284877aa234228ed81a24f: Tablet failed to bootstrap: IO error: Could 
not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 
/var/lib/kudu/tserver/wals/1bb9b2f91c3f48d7a97fb974112dedd6.recovery-1703662995452637:
 One or more errors occurredWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.504194 74 
ts_tablet_manager.cc:1356] T 1bb9b2f91c3f48d7a97fb974112dedd6 P 
ea0e0a381c284877aa234228ed81a24f: Time spent bootstrapping tablet: real 0.243s 
user 0.062s sys 0.046sWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.504212 74 
tablet_replica.cc:323] stopping tablet replicaWed, Dec 27 2023 3:43:15 pmI1227 
07:43:15.504217 74 raft_consensus.cc:2227] T 1bb9b2f91c3f48d7a97fb974112dedd6 P 
ea0e0a381c284877aa234228ed81a24f [term 1 FOLLOWER]: Raft consensus shutting 
down.Wed, Dec 27 2023 3:43:15 pmI1227 07:43:15.504230 74 
raft_consensus.cc:2256] T 1bb9b2f91c3f48d7a97fb974112dedd6 P 
ea0e0a381c284877aa234228ed81a24f [term 1 FOLLOWER]: Raft consensus is shut 
down!Wed, Dec 27 2023 3:43:15 pmI1227 07:43:15.504251 74 
tablet_bootstrap.cc:492] T d7eff00a19c44c728b4d46505c1ac5f2 P 
ea0e0a381c284877aa234228ed81a24f: Bootstrap starting.Wed, Dec 27 2023 3:43:15 
pmI1227 07:43:15.669176 74 tablet_bootstrap.cc:492] T 
d7eff00a19c44c728b4d46505c1ac5f2 P ea0e0a381c284877aa234228ed81a24f: Bootstrap 
replayed 1/1 log segments. Stats: ops{read=4975 overwritten=0 applied=4975 
ignored=0} inserts{seen=0 ignored=0} mutations{seen=0 ignored=0} 
orphaned_commits=0. Pending: 0 replicatesWed, Dec 27 2023 3:43:15 pmW1227 
07:43:15.687026 74 env_posix.cc:2337] Could not delete directory: IO error: 
/var/lib/kudu/tserver/wals/d7eff00a19c44c728b4d46505c1ac5f2.recovery-1703662995669230:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmW1227 07:43:15.687069 
74 env_posix.cc:2063] Error running callback with file 
/var/lib/kudu/tserver/wals/d7eff00a19c44c728b4d46505c1ac5f2.recovery-1703662995669230
 during walk: IO error: 
/var/lib/kudu/tserver/wals/d7eff00a19c44c728b4d46505c1ac5f2.recovery-1703662995669230:
 Directory not empty (error 39)Wed, Dec 27 2023 3:43:15 pmE1227 07:43:15.722580 
74 ts_tablet_manager.cc:1378] T d7eff00a19c44c728b4d46505c1ac5f2 P 
ea0e0a381c284877aa234228ed81a24f: Tablet failed to bootstrap: IO error: Could 
not remove renamed recovery dir 
/var/lib/kudu/tserver/wals/d7eff00a19c44c728b4d46505c1ac5f2.recovery-1703662995669230:
 
/var/lib/kudu/tserver/wals/d7eff00a19c44c728b4d46505c1ac5f2.recovery-1703662995669230:
 One or more errors occurredWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.722630 74 
ts_tablet_manager.cc:1356] T d7eff00a19c44c728b4d46505c1ac5f2 P 
ea0e0a381c284877aa234228ed81a24f: Time spent bootstrapping tablet: real 0.218s 
user 0.073s sys 0.048sWed, Dec 27 2023 3:43:15 pmI1227 07:43:15.722642 74 
tablet_replica.cc:323] stopping tablet replicaWed, Dec 27 2023 3:43:15 pmI1227 
07:43:15.722648 74 raft_consensus.cc:2227] T d7eff00a19c44c728b4d46505c1ac5f2 P 
ea0e0a381c284877aa234228ed81a24f [term 2 FOLLOWER]: Raft consensus shutting 
down.Wed, Dec 27 2023 3:43:15 pmI1227 07:43:15.722656 74 
raft_consensus.cc:2256] T d7eff00a19c44c728b4d46505c1ac5f2 P 
ea0e0a381c284877aa234228ed81a24f [term 2 FOLLOWER]: Raft consensus is shut 
down! {code}
 

 

 

 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to