Details
-
Bug
-
Resolution: Fixed
-
Critical
-
7.1.0
-
7.1.0-1275
-
Untriaged
-
Centos 64-bit
-
1
-
Yes
-
Magma May 2022
Description
Steps To Recreate:
Crash Recovery Test
- Created a 4 node cluster
- Loaded 5 million Items (item size = 1024 bytes)
- Start doc ops (create:expiry).
- During doc ops, keep killing memcached on all nodes we do wait for bucket warmup before triggering next sigkill. after bucket warm up there is a sleep of 30 to 60 secs before triggering next sigkill)
Memcached Logs:
Found on Node 172.23.106.204
CRITICAL [(default) magma_4]WALIterator corruption checksum error:Corruption: WAL::LogRec::Check corruption. Logrec:0x7fef17692fde offset:33:4062 flushOffset:4096
|
QE-TEST:
guides/gradlew --refresh-dependencies testrunner -P jython=/opt/jython/bin/jython -P 'args=-i /tmp/testexec.7900.ini bucket_storage=magma,rerun=false,GROUP=P1;recovery,randomize_value=true,doc_size=256,bucket_eviction_policy=fullEviction,replicas=0,nodes_init=4,enable_dp=false,collect_pcaps=True,upgrade_version=7.1.0-1275 -t storage.magma.magma_crash_recovery.MagmaCrashTests.test_crash_during_recovery,num_items=5000000,doc_size=1024,sdk_timeout=60,doc_ops=create:expiry,replicas=0,GROUP=P1;recovery,multiplier=10'
|