Re: trouble showing cluster scalability for read performance
Hi Diane, On 17/07/14 06:19, Diane Griffith wrote: We have been struggling proving out linear read performance with our cassandra configuration, that it is horizontally scaling. Wondering if anyone has any suggestions for what minimal configuration and approach to use to demonstrate this. We were trying to go for a simple set up, so on the keyspace and/or column families we went with the following settings thinking it was the minimal to prove scaling: replication_factor set to 1, a RF of 1 means that any particular bit of data exists on exactly one node. So if you are testing read speed by reading the same data item again and again as fast as you can, then all the reads will be coming from the same one node, the one that has that data item on it. In this situation adding more nodes won't help. Maybe this isn't exactly how you are testing read speed, but perhaps you are doing something analogous? I suggest you explain how you are measuring read speed exactly. Ciao, Duncan. SimpleStrategy, default consistency level, default compaction strategy (size tiered), but compacted down to 1 sstable per cf on each node (versus using leveled compaction for read performance) *Read Performance Results:* 1 client thread - 2 nodes 1 node was seen but we couldn't show increased performance adding more nodes i.e 4 nodes ! 2 nodes 2 client threads - 2 nodes 1 node still was true but again we couldn't show increased performance adding more nodes i.e. 4 nodes ! 2 nodes 10 client threads - this time 2 nodes 1 node on performance numbers. 2 nodes suffered from larger reduce throughput than 1 node was showing. Where are we going wrong? How have others shown horizontal scaling for reads? Thanks, Diane
Re: trouble showing cluster scalability for read performance
Duncan, Thanks for that feedback. I'll give a bit more info and then ask some more questions. *Our Goal*: Not to produce the fastest read but show horizontal scaling. *Test procedure*: * Inserted 54M rows where one third of that represents a unique key, 18M keys. End result given our schema is the 54M rows becomes 72M rows in the column family as the control query load to use. * have a client that queries 100k records in configurable batches, set to 1k. And then it does 100 reps of queries. It doesn't do the same keys for each rep, it uses an offset and then it increases the keys to query. * We can adjust the hit rate, i.e. how many of the keys will be found but have been focused on 100% hit rate * we run the query where multiple clients can be spawned to do the same query cycle 100k keys but the offset is not different so each client will query the same keys. * We thought we should manually compact the tables down to 1 sstable on a given node for consistent results across different cluster sizes * We had set replication factor to 1 originally to not complicate things or impact initial write times even. We would assess rf later was our thought. Since we changed the keys getting queried it would have to hit additional nodes to get row data but for just 1 client thread (to get simplest path to show horizontal scaling, had a slight decrease of performance when going to 4 nodes from 2 nodes) Things seen off of given procedure and set up: 1. 1 client thread: 2 nodes do better than 1 node on the query test. But 4 nodes did not do better than 2. 2. 2 client threads: 2 nodes were still doing better than 1 node 3. 10 client threads: the times drastically suffered and 2 nodes were doing 1/2 the speed of 1 node but before 1 to 2 threads performed better on 2 nodes vs 1 node. There was a huge decrease in performance on 2 nodes and just a mild decrease on 1 node. Note: 50+ threads was also drastically falling apart. *Observations*: - compacting each node to 1 table did not seem to help as running 10 client threads on exploded sstables and 2 nodes was 2x better than the last 2 node 10 client test but still decreased performance from 1 to 2 threads query against compacted tables - I would see upwards to 10 read requests pending at times while 8 to 10 were processing when I did nodetool tpstats. - having key cache on or disabled did not seem to impact things noticeably with our current configuration . *Questions:* 1. can multiple threads read the same sstable at the same time? Does compacting down to 1 sstable (to get a given row into one sstable) add any benefit or actually hurt like limited testing has indicated currently? 2. given the above testing process, does it still make sense to adjust replication factor appropriately for cluster size (i.e. 1 for 1 node cluster, 2 for 2 node cluster, 3 for n size cluster). We assumed it was just the ability for threads to connect into a coordinator that would help but sounds like it can still block I'm going to try a limited test with changing replication factor. But if anyone has any input on compacting to 1 sstable benefit or detriment on just simple scalability test, how if at all does cassandra block on reading sstables, and if higher replication factors do indeed help produce reliable results it would be appreciated. I know part of our charter was keep it simple to produce the scalability proof but it does sound like replication factor is hurting us if the delay between clients for the same keys is not long enough given the fact we are not doing different offsets for each client thread. Thanks, Diane On Thu, Jul 17, 2014 at 3:53 AM, Duncan Sands duncan.sa...@gmail.com wrote: Hi Diane, On 17/07/14 06:19, Diane Griffith wrote: We have been struggling proving out linear read performance with our cassandra configuration, that it is horizontally scaling. Wondering if anyone has any suggestions for what minimal configuration and approach to use to demonstrate this. We were trying to go for a simple set up, so on the keyspace and/or column families we went with the following settings thinking it was the minimal to prove scaling: replication_factor set to 1, a RF of 1 means that any particular bit of data exists on exactly one node. So if you are testing read speed by reading the same data item again and again as fast as you can, then all the reads will be coming from the same one node, the one that has that data item on it. In this situation adding more nodes won't help. Maybe this isn't exactly how you are testing read speed, but perhaps you are doing something analogous? I suggest you explain how you are measuring read speed exactly. Ciao, Duncan. SimpleStrategy, default consistency level, default compaction strategy (size tiered), but compacted down to 1 sstable per cf on each node (versus using leveled compaction for read performance)
Re: trouble showing cluster scalability for read performance
It sounds as if you are actually testing “vertical scalability” (load on a single node) rather than Cassandra’s sweet spot of “horizontal scalability” (add more nodes to handle higher load.) Maybe you could clarify your intentions and specific use case. Also, it sounds like you are trying to focus on large queries, but Cassandra’s sweet spot is lots of smaller queries. With larger queries you can end up measuring things like the capabilities of your hardware, cpu cores, memory, I/O bandwidth, network latency, JVM configuration, etc. rather than measuring Cassandra per se. So, again, maybe you could clarify your intended use case. It might be that you need to add more “vertical scale” (bigger box, more cores, more memory, beefier I/O and networking) to handle large queries, or maybe simple, Cassandra-style “horizontal scaling” (adding nodes) will be sufficient. Sure, you can tune Cassandra for single-node performance, but that seems lot a lot of extra work, to me, compared to adding more cheap nodes. -- Jack Krupansky From: Diane Griffith Sent: Thursday, July 17, 2014 9:31 AM To: user Subject: Re: trouble showing cluster scalability for read performance Duncan, Thanks for that feedback. I'll give a bit more info and then ask some more questions. Our Goal: Not to produce the fastest read but show horizontal scaling. Test procedure: * Inserted 54M rows where one third of that represents a unique key, 18M keys. End result given our schema is the 54M rows becomes 72M rows in the column family as the control query load to use. * have a client that queries 100k records in configurable batches, set to 1k. And then it does 100 reps of queries. It doesn't do the same keys for each rep, it uses an offset and then it increases the keys to query. * We can adjust the hit rate, i.e. how many of the keys will be found but have been focused on 100% hit rate * we run the query where multiple clients can be spawned to do the same query cycle 100k keys but the offset is not different so each client will query the same keys. * We thought we should manually compact the tables down to 1 sstable on a given node for consistent results across different cluster sizes * We had set replication factor to 1 originally to not complicate things or impact initial write times even. We would assess rf later was our thought. Since we changed the keys getting queried it would have to hit additional nodes to get row data but for just 1 client thread (to get simplest path to show horizontal scaling, had a slight decrease of performance when going to 4 nodes from 2 nodes) Things seen off of given procedure and set up: 1.. 1 client thread: 2 nodes do better than 1 node on the query test. But 4 nodes did not do better than 2. 2.. 2 client threads: 2 nodes were still doing better than 1 node 3.. 10 client threads: the times drastically suffered and 2 nodes were doing 1/2 the speed of 1 node but before 1 to 2 threads performed better on 2 nodes vs 1 node. There was a huge decrease in performance on 2 nodes and just a mild decrease on 1 node. Note: 50+ threads was also drastically falling apart. Observations: a.. compacting each node to 1 table did not seem to help as running 10 client threads on exploded sstables and 2 nodes was 2x better than the last 2 node 10 client test but still decreased performance from 1 to 2 threads query against compacted tables b.. I would see upwards to 10 read requests pending at times while 8 to 10 were processing when I did nodetool tpstats. c.. having key cache on or disabled did not seem to impact things noticeably with our current configuration . Questions: 1.. can multiple threads read the same sstable at the same time? Does compacting down to 1 sstable (to get a given row into one sstable) add any benefit or actually hurt like limited testing has indicated currently? 2.. given the above testing process, does it still make sense to adjust replication factor appropriately for cluster size (i.e. 1 for 1 node cluster, 2 for 2 node cluster, 3 for n size cluster). We assumed it was just the ability for threads to connect into a coordinator that would help but sounds like it can still block I'm going to try a limited test with changing replication factor. But if anyone has any input on compacting to 1 sstable benefit or detriment on just simple scalability test, how if at all does cassandra block on reading sstables, and if higher replication factors do indeed help produce reliable results it would be appreciated. I know part of our charter was keep it simple to produce the scalability proof but it does sound like replication factor is hurting us if the delay between clients for the same keys is not long enough given the fact we are not doing different offsets for each client thread. Thanks, Diane On Thu, Jul 17, 2014 at 3:53 AM, Duncan Sands duncan.sa...@gmail.com wrote: Hi Diane, On 17/07/14 06:19
Re: trouble showing cluster scalability for read performance
Definitely not trying to show vertical scaling. We have a query use case we are trying to show will scale as we add more nodes should performance fall below adequate. But to show the scaling we do the test on a 1 node cluster, then 2 node cluster, then 4 node cluster with a goal that query throughput increases when adding more nodes. Basically we do not want to tune for single node performance and did want to prove out adding nodes works but for our query use case it hasn't yet. Our query size is a valid use case though for our need. Earlier it may not have been clear but we are not querying the same key over and over in one thread but continuously querying random non duplicating keys. Bringing up the threading was not our main path or desired goal so I re-posted with clearer intent hopefully of our goal, what we experienced in the past against THRIFT and an older version of Cassandra which we have not been able to duplicate via CQL and Cassandra 2.0.6. So just hoping someone has suggestions of what one must do at a minimum to prove horizontal scaling or have suggestions of what to look at in our current datasize/query use case that may be causing us to not achieve horizontal scaling. Thanks, Diane On Thu, Jul 17, 2014 at 10:03 AM, Jack Krupansky j...@basetechnology.com wrote: It sounds as if you are actually testing “vertical scalability” (load on a single node) rather than Cassandra’s sweet spot of “horizontal scalability” (add more nodes to handle higher load.) Maybe you could clarify your intentions and specific use case. Also, it sounds like you are trying to focus on large queries, but Cassandra’s sweet spot is lots of smaller queries. With larger queries you can end up measuring things like the capabilities of your hardware, cpu cores, memory, I/O bandwidth, network latency, JVM configuration, etc. rather than measuring Cassandra per se. So, again, maybe you could clarify your intended use case. It might be that you need to add more “vertical scale” (bigger box, more cores, more memory, beefier I/O and networking) to handle large queries, or maybe simple, Cassandra-style “horizontal scaling” (adding nodes) will be sufficient. Sure, you can tune Cassandra for single-node performance, but that seems lot a lot of extra work, to me, compared to adding more cheap nodes. -- Jack Krupansky *From:* Diane Griffith dfgriff...@gmail.com *Sent:* Thursday, July 17, 2014 9:31 AM *To:* user user@cassandra.apache.org *Subject:* Re: trouble showing cluster scalability for read performance Duncan, Thanks for that feedback. I'll give a bit more info and then ask some more questions. *Our Goal*: Not to produce the fastest read but show horizontal scaling. *Test procedure*: * Inserted 54M rows where one third of that represents a unique key, 18M keys. End result given our schema is the 54M rows becomes 72M rows in the column family as the control query load to use. * have a client that queries 100k records in configurable batches, set to 1k. And then it does 100 reps of queries. It doesn't do the same keys for each rep, it uses an offset and then it increases the keys to query. * We can adjust the hit rate, i.e. how many of the keys will be found but have been focused on 100% hit rate * we run the query where multiple clients can be spawned to do the same query cycle 100k keys but the offset is not different so each client will query the same keys. * We thought we should manually compact the tables down to 1 sstable on a given node for consistent results across different cluster sizes * We had set replication factor to 1 originally to not complicate things or impact initial write times even. We would assess rf later was our thought. Since we changed the keys getting queried it would have to hit additional nodes to get row data but for just 1 client thread (to get simplest path to show horizontal scaling, had a slight decrease of performance when going to 4 nodes from 2 nodes) Things seen off of given procedure and set up: 1. 1 client thread: 2 nodes do better than 1 node on the query test. But 4 nodes did not do better than 2. 2. 2 client threads: 2 nodes were still doing better than 1 node 3. 10 client threads: the times drastically suffered and 2 nodes were doing 1/2 the speed of 1 node but before 1 to 2 threads performed better on 2 nodes vs 1 node. There was a huge decrease in performance on 2 nodes and just a mild decrease on 1 node. Note: 50+ threads was also drastically falling apart. *Observations*: - compacting each node to 1 table did not seem to help as running 10 client threads on exploded sstables and 2 nodes was 2x better than the last 2 node 10 client test but still decreased performance from 1 to 2 threads query against compacted tables - I would see upwards to 10 read requests pending at times while 8 to 10 were processing when I did nodetool tpstats
Re: trouble showing cluster scalability for read performance
Hi Diane, Sounds a bit like the client might be the limiting factor in your test - not the server. Especially if you're using one single threaded client, you might not be loading the backend in any significant way. Have you done any vertical scaling tests (identical client, bigger server)? if the client is indeed the limiting factor, then adding server capacity probably doesn't gain you much. What sort of CPU/IO load do you have on the client/server during your tests? I might be barking up the wrong tree (we haven't done any load tests yet on Cassandra), but when we load tested our clustered app, we used 3-10 client machines (with multithreaded clients) against 3 app server nodes. I would definitely first try to add more client load (multiple clients/multithreading and/or client machines) and once you're actually hitting the server properly, then add more server nodes. Best regards, Timo On 17 July 2014 20:39, Diane Griffith dfgriff...@gmail.com wrote: Definitely not trying to show vertical scaling. We have a query use case we are trying to show will scale as we add more nodes should performance fall below adequate. But to show the scaling we do the test on a 1 node cluster, then 2 node cluster, then 4 node cluster with a goal that query throughput increases when adding more nodes. Basically we do not want to tune for single node performance and did want to prove out adding nodes works but for our query use case it hasn't yet. Our query size is a valid use case though for our need. Earlier it may not have been clear but we are not querying the same key over and over in one thread but continuously querying random non duplicating keys. Bringing up the threading was not our main path or desired goal so I re-posted with clearer intent hopefully of our goal, what we experienced in the past against THRIFT and an older version of Cassandra which we have not been able to duplicate via CQL and Cassandra 2.0.6. So just hoping someone has suggestions of what one must do at a minimum to prove horizontal scaling or have suggestions of what to look at in our current datasize/query use case that may be causing us to not achieve horizontal scaling. Thanks, Diane On Thu, Jul 17, 2014 at 10:03 AM, Jack Krupansky j...@basetechnology.com wrote: It sounds as if you are actually testing “vertical scalability” (load on a single node) rather than Cassandra’s sweet spot of “horizontal scalability” (add more nodes to handle higher load.) Maybe you could clarify your intentions and specific use case. Also, it sounds like you are trying to focus on large queries, but Cassandra’s sweet spot is lots of smaller queries. With larger queries you can end up measuring things like the capabilities of your hardware, cpu cores, memory, I/O bandwidth, network latency, JVM configuration, etc. rather than measuring Cassandra per se. So, again, maybe you could clarify your intended use case. It might be that you need to add more “vertical scale” (bigger box, more cores, more memory, beefier I/O and networking) to handle large queries, or maybe simple, Cassandra-style “horizontal scaling” (adding nodes) will be sufficient. Sure, you can tune Cassandra for single-node performance, but that seems lot a lot of extra work, to me, compared to adding more cheap nodes. -- Jack Krupansky *From:* Diane Griffith dfgriff...@gmail.com *Sent:* Thursday, July 17, 2014 9:31 AM *To:* user user@cassandra.apache.org *Subject:* Re: trouble showing cluster scalability for read performance Duncan, Thanks for that feedback. I'll give a bit more info and then ask some more questions. *Our Goal*: Not to produce the fastest read but show horizontal scaling. *Test procedure*: * Inserted 54M rows where one third of that represents a unique key, 18M keys. End result given our schema is the 54M rows becomes 72M rows in the column family as the control query load to use. * have a client that queries 100k records in configurable batches, set to 1k. And then it does 100 reps of queries. It doesn't do the same keys for each rep, it uses an offset and then it increases the keys to query. * We can adjust the hit rate, i.e. how many of the keys will be found but have been focused on 100% hit rate * we run the query where multiple clients can be spawned to do the same query cycle 100k keys but the offset is not different so each client will query the same keys. * We thought we should manually compact the tables down to 1 sstable on a given node for consistent results across different cluster sizes * We had set replication factor to 1 originally to not complicate things or impact initial write times even. We would assess rf later was our thought. Since we changed the keys getting queried it would have to hit additional nodes to get row data but for just 1 client thread (to get simplest path to show horizontal scaling, had a slight decrease of performance when going to 4 nodes
trouble showing cluster scalability for read performance
We have been struggling proving out linear read performance with our cassandra configuration, that it is horizontally scaling. Wondering if anyone has any suggestions for what minimal configuration and approach to use to demonstrate this. We were trying to go for a simple set up, so on the keyspace and/or column families we went with the following settings thinking it was the minimal to prove scaling: replication_factor set to 1, SimpleStrategy, default consistency level, default compaction strategy (size tiered), but compacted down to 1 sstable per cf on each node (versus using leveled compaction for read performance) *Read Performance Results:* 1 client thread - 2 nodes 1 node was seen but we couldn't show increased performance adding more nodes i.e 4 nodes ! 2 nodes 2 client threads - 2 nodes 1 node still was true but again we couldn't show increased performance adding more nodes i.e. 4 nodes ! 2 nodes 10 client threads - this time 2 nodes 1 node on performance numbers. 2 nodes suffered from larger reduce throughput than 1 node was showing. Where are we going wrong? How have others shown horizontal scaling for reads? Thanks, Diane
