-
Type:
Bug
-
Status: Resolved
-
Priority:
Critical
-
Resolution: Fixed
-
Affects Version/s: 7.10
-
Component/s: Redis
-
Backlog priority:1,000
-
Sprint:nxFG 8.10.2
When all the nodes in a Redis cluster are restarted while Nuxeo is keeping up, some errors will be raised like
ERROR [http-bio-10.201.75.17-8080-exec-6] [nuxeo-error-log] redis.clients.jedis.exceptions.JedisDataException: NOSCRIPT No matching script. Please use EVAL. at redis.clients.jedis.Protocol.processError(Protocol.java:117) at redis.clients.jedis.Protocol.process(Protocol.java:142) at redis.clients.jedis.Protocol.read(Protocol.java:196) at redis.clients.jedis.Connection.readProtocolWithCheckingBroken(Connection.java:288) at redis.clients.jedis.Connection.getOne(Connection.java:270) at redis.clients.jedis.BinaryJedis.evalsha(BinaryJedis.java:3001) at redis.clients.jedis.BinaryJedis.evalsha(BinaryJedis.java:2994) at org.nuxeo.ecm.core.redis.contribs.RedisWorkQueuing$3.call(RedisWorkQueuing.java:513) at org.nuxeo.ecm.core.redis.contribs.RedisWorkQueuing$3.call(RedisWorkQueuing.java:509) at org.nuxeo.ecm.core.redis.RedisPoolExecutor.execute(RedisPoolExecutor.java:46) at org.nuxeo.ecm.core.redis.RedisFailoverExecutor$1.retry(RedisFailoverExecutor.java:39) at org.nuxeo.ecm.core.redis.retry.Retry.retry(Retry.java:54) at org.nuxeo.ecm.core.redis.RedisFailoverExecutor.execute(RedisFailoverExecutor.java:34) at org.nuxeo.ecm.core.redis.contribs.RedisWorkQueuing.addScheduledWork(RedisWorkQueuing.java:509) at org.nuxeo.ecm.core.redis.contribs.RedisBlockingQueue.putElement(RedisBlockingQueue.java:124) at org.nuxeo.ecm.core.work.NuxeoBlockingQueue.put(NuxeoBlockingQueue.java:85) at org.nuxeo.ecm.core.work.NuxeoBlockingQueue.offer(NuxeoBlockingQueue.java:68) at org.nuxeo.ecm.core.work.NuxeoBlockingQueue.offer(NuxeoBlockingQueue.java:33) at org.nuxeo.ecm.core.redis.contribs.RedisWorkQueuing.workSchedule(RedisWorkQueuing.java:182) at org.nuxeo.ecm.core.work.WorkManagerImpl$WorkThreadPoolExecutor.submit(WorkManagerImpl.java:545) at org.nuxeo.ecm.core.work.WorkManagerImpl$WorkThreadPoolExecutor.execute(WorkManagerImpl.java:534) at org.nuxeo.ecm.core.work.WorkManagerImpl.schedule(WorkManagerImpl.java:706) at org.nuxeo.ecm.core.work.WorkManagerImpl.schedule(WorkManagerImpl.java:651) at org.nuxeo.ecm.core.event.impl.AsyncEventExecutor.run(AsyncEventExecutor.java:78) at org.nuxeo.ecm.core.event.impl.EventServiceImpl.fireEventBundle(EventServiceImpl.java:302) at org.nuxeo.ecm.core.event.impl.EventServiceImpl.fireEvent(EventServiceImpl.java:244) at org.nuxeo.ecm.platform.ui.web.auth.NuxeoAuthenticationFilter.sendAuthenticationEvent(NuxeoAuthenticationFilter.java:200) at org.nuxeo.ecm.platform.ui.web.auth.NuxeoAuthenticationFilter.logAuthenticationAttempt(NuxeoAuthenticationFilter.java:233) at org.nuxeo.ecm.platform.ui.web.auth.NuxeoAuthenticationFilter.doAuthenticate(NuxeoAuthenticationFilter.java:291) at org.nuxeo.ecm.platform.ui.web.auth.NuxeoAuthenticationFilter.doFilterInternal(NuxeoAuthenticationFilter.java:532) at org.nuxeo.ecm.platform.ui.web.auth.service.NuxeoAuthFilterChain.doFilter(NuxeoAuthFilterChain.java:36) at org.nuxeo.ecm.platform.ui.web.auth.oauth.NuxeoOAuthFilter.doFilter(NuxeoOAuthFilter.java:120)
This seems to be due to the fact that some scripts are loaded to Redis memory when Nuxeo is initialized and these scripts are lost when a Redis node is restarted.
