You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've had great success using your project to benchmark Influxdb for [inserts] and wanted to confirm the details of its operation and see if their was a way to get the run phase working.
When I "run" workloads I get the error below, Using "load" I am able to insert 10mill records at about 12krps per without any stability issues. For comparison YCSB (not TS) "run" has worked fine on the same box.
Due to the descriptions stating they do [SCAN] & [READ] I had assumed all of the ops ran in the load phase. However only [INSERT] operations are logged.
The below error occurs regardless of the capacity of the server, I've used a 32 core/64gb and single core/8gb instance with xms-xmx configured correctly. It appears to occur under no load.
Starting test.
2016-11-01 14:52:36:337 0 sec: 0 operations; est completion in 0 seconds
2016-11-01 14:52:36:337 0 sec: 0 operations; est completion in 0 seconds
java.lang.NullPointerException
java.lang.NullPointerException
at com.yahoo.ycsb.db.InfluxDBClient.read(InfluxDBClient.java:140)
at com.yahoo.ycsb.DBWrapper.read(DBWrapper.java:85)
at com.yahoo.ycsb.workloads.CoreWorkload.doTransactionRead(CoreWorkload.java:990)
at com.yahoo.ycsb.workloads.CoreWorkload.doTransaction(CoreWorkload.java:946)
at com.yahoo.ycsb.ClientThread.run(Client.java:329)
java.lang.NullPointerException
at com.yahoo.ycsb.db.InfluxDBClient.read(InfluxDBClient.java:140)
at com.yahoo.ycsb.DBWrapper.read(DBWrapper.java:85)
at com.yahoo.ycsb.workloads.CoreWorkload.doTransactionRead(CoreWorkload.java:990)
at com.yahoo.ycsb.workloads.CoreWorkload.doTransaction(CoreWorkload.java:946)
at com.yahoo.ycsb.ClientThread.run(Client.java:329)
I also wanted to verify the accuracy of the description logged when running load operations as it states that scan and read ops are also done however the data only shows [INSERT] & [CLEANUP].
Workload A
[DESCRIPTION], Inserts 1000000 records with 3 tags each. Each tag has a value with a length of 10. Each record has a double value between 0 and 10000. After all inserts has processed there will be 1000 (random) read operations. Read will only use timestamp, no tags. All data will use 10000 predefined, randomly generated, tags. There is only one record per timestamp.
Workload B
[DESCRIPTION], Inserts 1000000 records with 3 tags each. Each tag has a value with a length of 10. Each record has a double value between 0 and 10000. After all inserts has processed there will be 100 sum/count/avg/scan operations (25 each) on a random timespan (min 1000 s, max 10000 s) with a random amount (between 1-5) of tag values per tag. All data will use 10000 predefined, randomly generated, tags. There is only one record per timestamp.
The text was updated successfully, but these errors were encountered:
I've had great success using your project to benchmark Influxdb for [inserts] and wanted to confirm the details of its operation and see if their was a way to get the run phase working.
When I "run" workloads I get the error below, Using "load" I am able to insert 10mill records at about 12krps per without any stability issues. For comparison YCSB (not TS) "run" has worked fine on the same box.
Due to the descriptions stating they do [SCAN] & [READ] I had assumed all of the ops ran in the load phase. However only [INSERT] operations are logged.
The below error occurs regardless of the capacity of the server, I've used a 32 core/64gb and single core/8gb instance with xms-xmx configured correctly. It appears to occur under no load.
I also wanted to verify the accuracy of the description logged when running load operations as it states that scan and read ops are also done however the data only shows [INSERT] & [CLEANUP].
Workload A
[DESCRIPTION], Inserts 1000000 records with 3 tags each. Each tag has a value with a length of 10. Each record has a double value between 0 and 10000. After all inserts has processed there will be 1000 (random) read operations. Read will only use timestamp, no tags. All data will use 10000 predefined, randomly generated, tags. There is only one record per timestamp.
Workload B
[DESCRIPTION], Inserts 1000000 records with 3 tags each. Each tag has a value with a length of 10. Each record has a double value between 0 and 10000. After all inserts has processed there will be 100 sum/count/avg/scan operations (25 each) on a random timespan (min 1000 s, max 10000 s) with a random amount (between 1-5) of tag values per tag. All data will use 10000 predefined, randomly generated, tags. There is only one record per timestamp.
The text was updated successfully, but these errors were encountered: