Hi Sachin, I guess these "few days of topology run" will be a week? Your Kafka spout is probably lagging behind and messages located at the offset requested were deleted from the Kafka topic (log.retention.hours), causing the following fetch request to throw `OffsetOutOfRange`.
Thanks, Guy On Mon, Nov 23, 2015 at 7:25 AM, Sachin Pasalkar < sachin_pasal...@symantec.com> wrote: > Can someone help us on this? > > From: Sachin Pasalkar <sachin_pasal...@symantec.com<mailto: > sachin_pasal...@symantec.com>> > Reply-To: "dev@storm.apache.org<mailto:dev@storm.apache.org>" < > dev@storm.apache.org<mailto:dev@storm.apache.org>> > Date: Friday, 20 November 2015 11:53 am > To: "dev@storm.apache.org<mailto:dev@storm.apache.org>" < > dev@storm.apache.org<mailto:dev@storm.apache.org>> > Subject: Why am I getting OffsetOutOfRange: Updating offset from offset? > > Hi, > > We are developing application where after some days of topology run, we > get continuous warning messages > > > 2015-11-20 05:05:42.226 s.k.KafkaUtils [WARN] Got fetch request with > offset out of range: [7238824446] > > 2015-11-20 05:05:42.229 s.k.t.TridentKafkaEmitter [WARN] OffsetOutOfRange: > Updating offset from offset = 7238824446 to offset = 7241183683 > > 2015-11-20 05:05:43.207 s.k.KafkaUtils [WARN] Got fetch request with > offset out of range: [7022945051] > > 2015-11-20 05:05:43.208 s.k.t.TridentKafkaEmitter [WARN] OffsetOutOfRange: > Updating offset from offset = 7022945051 to offset = 7025309343 > > 2015-11-20 05:05:44.260 s.k.KafkaUtils [WARN] Got fetch request with > offset out of range: [7170559432] > > 2015-11-20 05:05:44.264 s.k.t.TridentKafkaEmitter [WARN] OffsetOutOfRange: > Updating offset from offset = 7170559432 to offset = 7172920769 > > 2015-11-20 05:05:45.332 s.k.KafkaUtils [WARN] Got fetch request with > offset out of range: [7132495867]…… > > > After some point topology stop processing messages, I need to rebalance it > to start it again. > > > My spout config is > > > BrokerHosts brokers = new ZkHosts((String) > stormConfiguration.get(ZOOKEEPER_HOSTS)); > > TridentKafkaConfig spoutConfig = new TridentKafkaConfig(brokers, (String) > stormConfiguration.get(KAFKA_INPUT_TOPIC)); > > > spoutConfig.scheme = getSpoutScheme(stormConfiguration); > > Boolean forceFromStart = (Boolean) > stormConfiguration.get(FORCE_FROM_START); > > > spoutConfig.ignoreZkOffsets = false; > > spoutConfig.fetchSizeBytes = > stormConfiguration.getIntProperty(KAFKA_CONSUMER_FETCH_SIZE_BYTE, > KAFKA_CONSUMER_DEFAULT_FETCH_SIZE_BYTE); > > spoutConfig.bufferSizeBytes = > stormConfiguration.getIntProperty(KAFKA_CONSUMER_BUFFER_SIZE_BYTE, > KAFKA_CONSUMER_DEFAULT_BUFFER_SIZE_BYTE); > > As per my knowledge, only thing we are doing wrong is topic has 12 > partitions but we are reading using only 1 spout, but that’s limitation on > our side. I am not sure why its getting halted? It just keep printing below > lines and does nothing > > > 2015-11-20 05:44:41.574 b.s.m.n.Server [INFO] Getting metrics for server > on port 6700 > > 2015-11-20 05:44:41.574 b.s.m.n.Client [INFO] Getting metrics for client > connection to Netty-Client-b-bdata-xx.net/xxx.xx.xxx.xxx:6700 > > 2015-11-20 05:44:41.574 b.s.m.n.Client [INFO] Getting metrics for client > connection to Netty-Client-b-bdata-xx.net/xxx.xx.xxx.xxx:6709 > > 2015-11-20 05:44:41.574 b.s.m.n.Client [INFO] Getting metrics for client > connection to Netty-Client-b-bdata-xx.net/xxx.xx.xxx.xxx:6707 > > > Thanks, > > Sachin > >
