* parallel transaction submit @ 2016-08-25 7:45 Tang, Haodong 2016-08-25 7:55 ` Haomai Wang 0 siblings, 1 reply; 6+ messages in thread From: Tang, Haodong @ 2016-08-25 7:45 UTC (permalink / raw) To: sweil, varada.kari, ceph-devel Hi Sage, Varada Noticed you are making parallel transaction submits, we also worked out a prototype that looks similar, here is the link for the implementation: https://github.com/ceph/ceph/pull/10856 Background: From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. Implementation: The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. Performance evaluation: Test ENV: 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. Performance: We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. Test summary: QD Scaling Test - 4k Random Write: QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 With rocksdb (IOPS) 682 173000 190000 203000 204000 With memdb (IOPS) 704 180000 194000 206000 218000 With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 It seems single thread of transaction submits will be a bottleneck if using MemDB. ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: parallel transaction submit 2016-08-25 7:45 parallel transaction submit Tang, Haodong @ 2016-08-25 7:55 ` Haomai Wang 2016-08-25 8:47 ` Varada Kari 2016-08-25 8:48 ` Tang, Haodong 0 siblings, 2 replies; 6+ messages in thread From: Haomai Wang @ 2016-08-25 7:55 UTC (permalink / raw) To: Tang, Haodong; +Cc: sweil, varada.kari, ceph-devel looks very litlle improvements. rocksdb result meet my expectation because rocksdb internal has lock for multi sync write. But memdb improments is a little confusing. On Thu, Aug 25, 2016 at 3:45 PM, Tang, Haodong <haodong.tang@intel.com> wrote: > Hi Sage, Varada > > Noticed you are making parallel transaction submits, we also worked out a prototype that looks similar, here is the link for the implementation: https://github.com/ceph/ceph/pull/10856 > > Background: > From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. > > Implementation: > The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. > > Performance evaluation: > Test ENV: > 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. > Performance: > We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. > > What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. > > Test summary: > QD Scaling Test - 4k Random Write: > QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 > With rocksdb (IOPS) 682 173000 190000 203000 204000 > With memdb (IOPS) 704 180000 194000 206000 218000 > With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 > With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 > > > It seems single thread of transaction submits will be a bottleneck if using MemDB. > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: parallel transaction submit 2016-08-25 7:55 ` Haomai Wang @ 2016-08-25 8:47 ` Varada Kari 2016-08-25 14:11 ` Sage Weil 2016-08-25 8:48 ` Tang, Haodong 1 sibling, 1 reply; 6+ messages in thread From: Varada Kari @ 2016-08-25 8:47 UTC (permalink / raw) To: Haomai Wang, Tang, Haodong; +Cc: sweil, ceph-devel Hi, Increasing the number of the kv_sync_threads is not giving much of performance. In the current threading model, shard_worker submits the IO to the block device which are handled by aio_callback thread(which is one) and submits to the kv_sync thread, which batches the requests and submits to the rocksdb. Because kv_sync batches the requests and submits the requests, we might observe more time spent on kv_sync_thread routine. And i haven't observed much of an improvement by adding more threads here. But when increased the number of callbacks thread from aio(still needs some refinements in polling for the request completions) and completing the write completion in the same thread context increased some performance. I don't have the numbers to say how much, but that is better than having multiple kv_sync threads, adding one more queue and lock. You can refer to https://github.com/varadakari/ceph/commits/wip-parallel-aiocb (ignore the first commit, was trying to do sync transaction in the same thread context of sharded worker to measure the latency). was exploring a way to have the aio callback thread matached/reserved at the time of io submission, so that we don't need to do io_getevents(), kind of a async callback to the specified thread so that we can avoid some waiting logic in io_getevents() and process the request in the same thread context. You can refer to http://manpages.ubuntu.com/manpages/wily/man3/io_set_callback.3.html. I don't have the working code ready for this. FWIW, that is worth experimenting and see if it reduces any latency. Varada On Thursday 25 August 2016 01:25 PM, Haomai Wang wrote: > looks very litlle improvements. rocksdb result meet my expectation > because rocksdb internal has lock for multi sync write. But memdb > improments is a little confusing. > > On Thu, Aug 25, 2016 at 3:45 PM, Tang, Haodong <haodong.tang@intel.com> wrote: >> Hi Sage, Varada >> >> Noticed you are making parallel transaction submits, we also worked out a prototype that looks similar, here is the link for the implementation: https://github.com/ceph/ceph/pull/10856 >> >> Background: >> From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. >> >> Implementation: >> The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. >> >> Performance evaluation: >> Test ENV: >> 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. >> Performance: >> We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. >> >> What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. >> >> Test summary: >> QD Scaling Test - 4k Random Write: >> QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 >> With rocksdb (IOPS) 682 173000 190000 203000 204000 >> With memdb (IOPS) 704 180000 194000 206000 218000 >> With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 >> With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 >> >> >> It seems single thread of transaction submits will be a bottleneck if using MemDB. >> -- >> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html PLEASE NOTE: The information contained in this electronic mail message is intended only for the use of the designated recipient(s) named above. If the reader of this message is not the intended recipient, you are hereby notified that you have received this message in error and that any review, dissemination, distribution, or copying of this message is strictly prohibited. If you have received this communication in error, please notify the sender by telephone or e-mail (as shown above) immediately and destroy any and all copies of this message in your possession (whether hard copies or electronically stored copies). ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: parallel transaction submit 2016-08-25 8:47 ` Varada Kari @ 2016-08-25 14:11 ` Sage Weil 2016-08-25 14:26 ` Varada Kari 0 siblings, 1 reply; 6+ messages in thread From: Sage Weil @ 2016-08-25 14:11 UTC (permalink / raw) To: Varada Kari; +Cc: Haomai Wang, Tang, Haodong, ceph-devel On Thu, 25 Aug 2016, Varada Kari wrote: > Hi, > > Increasing the number of the kv_sync_threads is not giving much of > performance. In the current threading model, shard_worker submits the IO > to the block device which are handled by aio_callback thread(which is > one) and submits to the kv_sync thread, which batches the requests and > submits to the rocksdb. Because kv_sync batches the requests and submits > the requests, we might observe more time spent on kv_sync_thread > routine. And i haven't observed much of an improvement by adding more > threads here. > > But when increased the number of callbacks thread from aio(still needs > some refinements in polling for the request completions) and completing > the write completion in the same thread context increased some > performance. I don't have the numbers to say how much, but that is > better than having multiple kv_sync threads, adding one more queue and > lock. You can refer to > https://github.com/varadakari/ceph/commits/wip-parallel-aiocb (ignore > the first commit, was trying to do sync transaction in the same thread > context of sharded worker to measure the latency). Yeah, I think this is right. I see two avenues of attack: - Try to eliminate the handoff to _kv_sync_thread by having the transaction submitted to rocksdb in the calling thread. This will require a bit of refactoring but I think it's possible. We don't actually want to block, though, so it'll be an async submission, and we'll still need kv_sync_thread just telling rocksdb to commit in a loop and triggering callbacks. A recent PR sharded the completion finishers so I'm guessing the final step would be some affinity thing that pins the finishers to the same cores as the submitters? - Shard the io completion (before we submit the kv transaction). Not sure if we want a thread per shard, or polls at opportunistic/strategic points in code. The goal would be keeping the processing local to the core/socket (vs the current strategy of a single thread waiting/polling for completions and doing the next phase of work). > was exploring a way to have the aio callback thread matached/reserved at > the time of io submission, so that we don't need to do io_getevents(), > kind of a async callback to the specified thread so that we can avoid > some waiting logic in io_getevents() and process the request in the same > thread context. You can refer to > http://manpages.ubuntu.com/manpages/wily/man3/io_set_callback.3.html. I > don't have the working code ready for this. FWIW, that is worth > experimenting and see if it reduces any latency. I don't think this will help--it just means you're using a layer of the library that's calling getevents for you and calling your callback. Thanks! sage > > Varada > > On Thursday 25 August 2016 01:25 PM, Haomai Wang wrote: > > looks very litlle improvements. rocksdb result meet my expectation > > because rocksdb internal has lock for multi sync write. But memdb > > improments is a little confusing. > > > > On Thu, Aug 25, 2016 at 3:45 PM, Tang, Haodong <haodong.tang@intel.com> wrote: > >> Hi Sage, Varada > >> > >> Noticed you are making parallel transaction submits, we also worked out a prototype that looks similar, here is the link for the implementation: https://github.com/ceph/ceph/pull/10856 > >> > >> Background: > >> From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. > >> > >> Implementation: > >> The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. > >> > >> Performance evaluation: > >> Test ENV: > >> 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. > >> Performance: > >> We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. > >> > >> What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. > >> > >> Test summary: > >> QD Scaling Test - 4k Random Write: > >> QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 > >> With rocksdb (IOPS) 682 173000 190000 203000 204000 > >> With memdb (IOPS) 704 180000 194000 206000 218000 > >> With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 > >> With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 > >> > >> > >> It seems single thread of transaction submits will be a bottleneck if using MemDB. > >> -- > >> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > >> the body of a message to majordomo@vger.kernel.org > >> More majordomo info at http://vger.kernel.org/majordomo-info.html > > PLEASE NOTE: The information contained in this electronic mail message is intended only for the use of the designated recipient(s) named above. If the reader of this message is not the intended recipient, you are hereby notified that you have received this message in error and that any review, dissemination, distribution, or copying of this message is strictly prohibited. If you have received this communication in error, please notify the sender by telephone or e-mail (as shown above) immediately and destroy any and all copies of this message in your possession (whether hard copies or electronically stored copies). > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > > ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: parallel transaction submit 2016-08-25 14:11 ` Sage Weil @ 2016-08-25 14:26 ` Varada Kari 0 siblings, 0 replies; 6+ messages in thread From: Varada Kari @ 2016-08-25 14:26 UTC (permalink / raw) To: Sage Weil; +Cc: Haomai Wang, Tang, Haodong, ceph-devel On Thursday 25 August 2016 07:41 PM, Sage Weil wrote: > On Thu, 25 Aug 2016, Varada Kari wrote: >> Hi, >> >> Increasing the number of the kv_sync_threads is not giving much of >> performance. In the current threading model, shard_worker submits the IO >> to the block device which are handled by aio_callback thread(which is >> one) and submits to the kv_sync thread, which batches the requests and >> submits to the rocksdb. Because kv_sync batches the requests and submits >> the requests, we might observe more time spent on kv_sync_thread >> routine. And i haven't observed much of an improvement by adding more >> threads here. >> >> But when increased the number of callbacks thread from aio(still needs >> some refinements in polling for the request completions) and completing >> the write completion in the same thread context increased some >> performance. I don't have the numbers to say how much, but that is >> better than having multiple kv_sync threads, adding one more queue and >> lock. You can refer to >> https://github.com/varadakari/ceph/commits/wip-parallel-aiocb (ignore >> the first commit, was trying to do sync transaction in the same thread >> context of sharded worker to measure the latency). > Yeah, I think this is right. I see two avenues of attack: > > - Try to eliminate the handoff to _kv_sync_thread by having the > transaction submitted to rocksdb in the calling thread. This will > require a bit of refactoring but I think it's possible. We don't actually > want to block, though, so it'll be an async submission, and we'll still > need kv_sync_thread just telling rocksdb to commit in a loop and > triggering callbacks. A recent PR sharded the completion finishers so I'm > guessing the final step would be some affinity thing that pins the > finishers to the same cores as the submitters? I kind of copied what kv_sync_thread does and was able to run multiple callbacks at the same time. If we can complete the write ack in the same aio_cb thread not handing off to finisher thread, we can eliminate one thread switch and we can have same number of threads as shards. We can use the same logic to handle the finishers (not sure if we can process the request by reading the osr here) for the callback threads. Varada > - Shard the io completion (before we submit the kv transaction). Not sure > if we want a thread per shard, or polls at opportunistic/strategic points > in code. The goal would be keeping the processing local to the > core/socket (vs the current strategy of a single thread waiting/polling > for completions and doing the next phase of work). > >> was exploring a way to have the aio callback thread matached/reserved at >> the time of io submission, so that we don't need to do io_getevents(), >> kind of a async callback to the specified thread so that we can avoid >> some waiting logic in io_getevents() and process the request in the same >> thread context. You can refer to >> http://manpages.ubuntu.com/manpages/wily/man3/io_set_callback.3.html. I >> don't have the working code ready for this. FWIW, that is worth >> experimenting and see if it reduces any latency. > I don't think this will help--it just means you're using a layer of the > library that's calling getevents for you and calling your callback. > > Thanks! > sage > > >> Varada >> >> On Thursday 25 August 2016 01:25 PM, Haomai Wang wrote: >>> looks very litlle improvements. rocksdb result meet my expectation >>> because rocksdb internal has lock for multi sync write. But memdb >>> improments is a little confusing. >>> >>> On Thu, Aug 25, 2016 at 3:45 PM, Tang, Haodong <haodong.tang@intel.com> wrote: >>>> Hi Sage, Varada >>>> >>>> Noticed you are making parallel transaction submits, we also worked out a prototype that looks similar, here is the link for the implementation: https://github.com/ceph/ceph/pull/10856 >>>> >>>> Background: >>>> From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. >>>> >>>> Implementation: >>>> The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. >>>> >>>> Performance evaluation: >>>> Test ENV: >>>> 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. >>>> Performance: >>>> We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. >>>> >>>> What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. >>>> >>>> Test summary: >>>> QD Scaling Test - 4k Random Write: >>>> QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 >>>> With rocksdb (IOPS) 682 173000 190000 203000 204000 >>>> With memdb (IOPS) 704 180000 194000 206000 218000 >>>> With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 >>>> With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 >>>> >>>> >>>> It seems single thread of transaction submits will be a bottleneck if using MemDB. >>>> -- >>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in >>>> the body of a message to majordomo@vger.kernel.org >>>> More majordomo info at http://vger.kernel.org/majordomo-info.html >> PLEASE NOTE: The information contained in this electronic mail message is intended only for the use of the designated recipient(s) named above. If the reader of this message is not the intended recipient, you are hereby notified that you have received this message in error and that any review, dissemination, distribution, or copying of this message is strictly prohibited. If you have received this communication in error, please notify the sender by telephone or e-mail (as shown above) immediately and destroy any and all copies of this message in your possession (whether hard copies or electronically stored copies). >> -- >> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> >> ^ permalink raw reply [flat|nested] 6+ messages in thread
* RE: parallel transaction submit 2016-08-25 7:55 ` Haomai Wang 2016-08-25 8:47 ` Varada Kari @ 2016-08-25 8:48 ` Tang, Haodong 1 sibling, 0 replies; 6+ messages in thread From: Tang, Haodong @ 2016-08-25 8:48 UTC (permalink / raw) To: Haomai Wang; +Cc: sweil, varada.kari, ceph-devel Yeah, from the perf counter, we do reduce time waiting for kv_queue, but increase rocksdb submisstion time. Seems it is better to do batch submission for rocksdb instead of parallel submission. Besides, parallel submission also make bdev->flush more frequent. From the test result, there is a little performance improvement if using memdb, but compared with rocksdb, throughput is more stable. -----Original Message----- From: Haomai Wang [mailto:haomai@xsky.com] Sent: Thursday, August 25, 2016 3:55 PM To: Tang, Haodong <haodong.tang@intel.com> Cc: sweil@redhat.com; varada.kari@sandisk.com; ceph-devel@vger.kernel.org Subject: Re: parallel transaction submit looks very litlle improvements. rocksdb result meet my expectation because rocksdb internal has lock for multi sync write. But memdb improments is a little confusing. On Thu, Aug 25, 2016 at 3:45 PM, Tang, Haodong <haodong.tang@intel.com> wrote: > Hi Sage, Varada > > Noticed you are making parallel transaction submits, we also worked > out a prototype that looks similar, here is the link for the > implementation: https://github.com/ceph/ceph/pull/10856 > > Background: > From the perf counter we added, found it spent a lot time in kv_queue, that is, single thread transaction submits is not competent to handle the transaction from OSD. > > Implementation: > The key thought is to use multiple thread and assign each TransContext to one of the processing threads. In order to parallelize transaction submit, add different kv_locks and kv_conds for each thread. > > Performance evaluation: > Test ENV: > 4 x server, 4 x client, 16 x Intel S3700 as block device, and 4 x Intel P3600 as Rocksdb/WAL device. > Performance: > We also did several quick tests to verify the performance benefit, the results showed that parallel transaction submission will brought 10% performance improvement if using memdb, but little performance improvement with rocksdb. > > What's more, without parallel transaction submits, we also see performance boost if just changing to MemDB, but a little. > > Test summary: > QD Scaling Test - 4k Random Write: > QD = 1 QD = 16 QD = 32 QD = 64 QD = 128 > With rocksdb (IOPS) 682 173000 190000 203000 204000 > With memdb (IOPS) 704 180000 194000 206000 218000 > With rocksdb+multiple_kv_thread(IOPS) / 164243 167037 180961 201752 > With memdb+multiple_kv_thread(IOPS) / 176000 200000 221000 227000 > > > It seems single thread of transaction submits will be a bottleneck if using MemDB. > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" > in the body of a message to majordomo@vger.kernel.org More majordomo > info at http://vger.kernel.org/majordomo-info.html ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2016-08-25 14:42 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2016-08-25 7:45 parallel transaction submit Tang, Haodong 2016-08-25 7:55 ` Haomai Wang 2016-08-25 8:47 ` Varada Kari 2016-08-25 14:11 ` Sage Weil 2016-08-25 14:26 ` Varada Kari 2016-08-25 8:48 ` Tang, Haodong
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.