Thanks Chris, this sounds really “fishy”. Let me try to debug this a bit more (not sure yet how tho).
> On 20 Jul 2016, at 19:06, Chris Conroy <[email protected]> wrote: > > This leak resulted in exhaustion of 4G of NIO memory. The same application > under Netty 3 only ever uses <200MB of NIO memory. I have run several > experiments where I take some traffic, disable the node from serving traffic, > and then force a full GC. The allocated NIO memory does not return to normal > levels. > > Spot checking the byte buf handles from my heap dump, I see lots (all?) > `recycle` and `lastRecycleIds` of `int -2147483648 = 0x80000000` > > The experiment yesterday with the recycler disabled did not result in any > such leaks over several hours of traffic. > > On Wed, Jul 20, 2016 at 12:41 AM, 'Norman Maurer' via Netty discussions > <[email protected] <mailto:[email protected]>> wrote: > Are you actually sure these are really leaked ? > > The point of the “ThreadLocalUnsafeDirectByteBuf” is that it can be reused so > it is expected that it not gets released after its written but put back in > the recycler. Or you saying there are too many of these ? > > >> On 19 Jul 2016, at 22:39, Chris Conroy <[email protected] >> <mailto:[email protected]>> wrote: >> >> Yes all the leaked buffers look to be ThreadLocalUnsafeDirectByteBuf >> >> Disabling the recycler w/o backpressure on slow readers resulted in several >> massive allocation events. I was able to free them by forcing GC so no leaks >> there. Memory usage with backpressure on slow readers is back down to Netty >> 3 w/o backpressure levels. I don't see any leaks so far. >> >> Assuming that disabling the recycler does indeed prevent leaks, where do we >> go from there? Any ideas of things to look for in my application that might >> cause the recycler to get into a bad state? >> >> On Tue, Jul 19, 2016 at 3:40 PM, 'Norman Maurer' via Netty discussions >> <[email protected] <mailto:[email protected]>> wrote: >> Btw all the “leaking” buffers are always: >> >> ThreadLocalUnsafeDirectByteBuf >> >> ? >> >> >>> On 19 Jul 2016, at 21:29, Chris Conroy <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> I ran with paranoid while observing the leak and got no messages from the >>> ResourceLeakDetector. >>> >>> Initial results from recycler maxCapacity of 0 are looking positive with >>> respect to the leak. However, I did see a large spike of NIO memory >>> allocated (1G) where the Netty 3 version of this app receiving similar >>> traffic peaks at about 1/10th that. I'll let this run for a while and >>> report back later with the full results. >>> >>> On Tue, Jul 19, 2016 at 2:59 PM, 'Norman Maurer' via Netty discussions >>> <[email protected] <mailto:[email protected]>> wrote: >>> Can you do me a favour and run with: >>> >>> -Dio.netty.recycler.maxCapacity=0 >>> >>> And let me know if you still see some leaks ? >>> >>> >>> >>>> On 19 Jul 2016, at 20:50, Chris Conroy <[email protected] >>>> <mailto:[email protected]>> wrote: >>>> >>>> 4.1.0.Final >>>> >>>> On Tue, Jul 19, 2016 at 2:48 PM, 'Norman Maurer' via Netty discussions >>>> <[email protected] <mailto:[email protected]>> wrote: >>>> Are you using 4.0 or 4.1 ? >>>> >>>>> On 19 Jul 2016, at 20:42, Chris Conroy <[email protected] >>>>> <mailto:[email protected]>> wrote: >>>>> >>>>> I have not been able to reproduce locally yet, but I do see it in a >>>>> cluster that takes a lot of varied traffic. The leak detector has not >>>>> fired for this under advanced. I will give paranoid a shot to be safe, >>>>> but it's my understanding that the leak detection framework is more for >>>>> dealing with pooled byte buf misuse, but in this case I am exclusively >>>>> using unpooled heap byte bufs: these are just the socket direct byte bufs >>>>> that appear to be leaking. >>>>> >>>>> I meant to add this earlier: The path to GC root goes: >>>>> >>>>> io.netty.buffer.ByteBufUtil$ThreadLocalUnsafeDirectByteBuf >>>>> io.netty.util.Recycler$DefaultHandle >>>>> io.netty.util.Recycler$DefaultHandle[] >>>>> io.netty.util.Recycler#Stack >>>>> java.lang.Object[] >>>>> io.netty.util.internal.InternalThreadLocalMap >>>>> ... (more thread local map refs up to java.lang.Thread) >>>>> >>>>> >>>>> >>>>> On Tuesday, July 19, 2016 at 2:17:25 PM UTC-4, Norman Maurer wrote: >>>>> Can you provide a reproducer? Also did you try to run with paranoid leak >>>>> detection >>>>> >>>>> Am 19.07.2016 um 20:04 schrieb Chris Conroy <[email protected] <>>: >>>>> >>>>>> Ah okay: I didn't see the calls to failFlushed since they occur above >>>>>> the stanza I found suspicious. >>>>>> >>>>>> So, the above explanation is probably not correct. Still, I am seeing a >>>>>> leak where DirectByteBufs are rooted to the recycler, and the speed at >>>>>> which these buffers leak appears to be correlated with slow/partial >>>>>> readers. >>>>>> >>>>>> On Monday, July 18, 2016 at 4:36:31 PM UTC-4, Norman Maurer wrote: >>>>>> failFlushed(...) should be called to fail and release all flushed >>>>>> messages. >>>>>> >>>>>> Are you saying this not happens? >>>>>> >>>>>> Am 18.07.2016 um 22:02 schrieb Chris Conroy <[email protected] <>>: >>>>>> >>>>>>> I’ve been trying to track down a NIO memory leak that occurs in a Netty >>>>>>> application I am porting from Netty 3 to Netty 4. This leak does not >>>>>>> occur in the Netty 3 version of the application. >>>>>>> >>>>>>> For now, I’m using only unpooled heap buffers in Netty 4, but NIO >>>>>>> buffers do come into play for socket communication. >>>>>>> >>>>>>> I’ve captured a few heap dumps from affected instances, and in each it >>>>>>> appears that the leaked DirectByteBuf java objects are rooted in an >>>>>>> io.netty.util.Recycler. >>>>>>> >>>>>>> These buffers remain indefinitely: I can disable the application to >>>>>>> drain traffic and force GCs, but the # of NIO buffers and NIO allocated >>>>>>> space stays flat. >>>>>>> >>>>>>> The issue is likely related to slow readers. However, the leak persists >>>>>>> long after all channels have been closed. >>>>>>> >>>>>>> I implemented a writability listener and the leak does appear to go >>>>>>> away if I stop writing to a channel after it goes unwritable. This is >>>>>>> good, but I’m still worried that this just makes the problem less >>>>>>> likely since it’s still possible to write/flush and have pending data: >>>>>>> writability just limits how much data will be buffered. >>>>>>> >>>>>>> Digging into ChannelOutBoundBuffer I see the following stanza in close: >>>>>>> >>>>>>> >>>>>>> // Release all unflushed messages. >>>>>>> try { >>>>>>> Entry e = unflushedEntry; >>>>>>> while (e != null) { >>>>>>> // Just decrease; do not trigger any events via >>>>>>> decrementPendingOutboundBytes() >>>>>>> int size = e.pendingSize; >>>>>>> TOTAL_PENDING_SIZE_UPDATER.addAndGet(this, -size); >>>>>>> >>>>>>> if (!e.cancelled) { >>>>>>> ReferenceCountUtil.safeRelease(e.msg); >>>>>>> safeFail(e.promise, cause); >>>>>>> } >>>>>>> e = e.recycleAndGetNext(); >>>>>>> } >>>>>>> } finally { >>>>>>> inFail = false; >>>>>>> } >>>>>>> clearNioBuffers(); >>>>>>> This seems a bit curious to me: why are flushed buffers not released >>>>>>> here? Since the leak seems to be rooted in the Recycler, this could be >>>>>>> the culprit…What do you think? >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> You received this message because you are subscribed to the Google >>>>>>> Groups "Netty discussions" group. >>>>>>> To unsubscribe from this group and stop receiving emails from it, send >>>>>>> an email to [email protected] <>. >>>>>>> To view this discussion on the web visit >>>>>>> https://groups.google.com/d/msgid/netty/CA%2B%3DgZKADssKFcs-WCc8%2Br2RWrvbgg3csaJPdcsXL_mCD5yG2bg%40mail.gmail.com >>>>>>> >>>>>>> <https://groups.google.com/d/msgid/netty/CA%2B%3DgZKADssKFcs-WCc8%2Br2RWrvbgg3csaJPdcsXL_mCD5yG2bg%40mail.gmail.com?utm_medium=email&utm_source=footer>. >>>>>>> For more options, visit https://groups.google.com/d/optout >>>>>>> <https://groups.google.com/d/optout>. >>>>>> >>>>>> >>>>>> -- >>>>>> You received this message because you are subscribed to the Google >>>>>> Groups "Netty discussions" group. >>>>>> To unsubscribe from this group and stop receiving emails from it, send >>>>>> an email to [email protected] <>. >>>>>> To view this discussion on the web visit >>>>>> https://groups.google.com/d/msgid/netty/b66894c3-1e65-4235-9201-b4f1dca11a81%40googlegroups.com >>>>>> >>>>>> <https://groups.google.com/d/msgid/netty/b66894c3-1e65-4235-9201-b4f1dca11a81%40googlegroups.com?utm_medium=email&utm_source=footer>. >>>>>> For more options, visit https://groups.google.com/d/optout >>>>>> <https://groups.google.com/d/optout>. >>>>> >>>>> >>>>> -- >>>>> You received this message because you are subscribed to the Google Groups >>>>> "Netty discussions" group. >>>>> To unsubscribe from this group and stop receiving emails from it, send an >>>>> email to [email protected] >>>>> <mailto:[email protected]>. >>>>> To view this discussion on the web visit >>>>> https://groups.google.com/d/msgid/netty/e823494b-caf1-4b1f-b629-405bbdbf4c40%40googlegroups.com >>>>> >>>>> <https://groups.google.com/d/msgid/netty/e823494b-caf1-4b1f-b629-405bbdbf4c40%40googlegroups.com?utm_medium=email&utm_source=footer>. >>>>> For more options, visit https://groups.google.com/d/optout >>>>> <https://groups.google.com/d/optout>. >>>> >>>> >>>> -- >>>> You received this message because you are subscribed to a topic in the >>>> Google Groups "Netty discussions" group. >>>> To unsubscribe from this topic, visit >>>> https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe >>>> <https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe>. >>>> To unsubscribe from this group and all its topics, send an email to >>>> [email protected] >>>> <mailto:[email protected]>. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/netty/F788648D-2C4E-4031-BD2A-EFFAEED64BDC%40googlemail.com >>>> >>>> <https://groups.google.com/d/msgid/netty/F788648D-2C4E-4031-BD2A-EFFAEED64BDC%40googlemail.com?utm_medium=email&utm_source=footer>. >>>> >>>> For more options, visit https://groups.google.com/d/optout >>>> <https://groups.google.com/d/optout>. >>>> >>>> >>>> -- >>>> You received this message because you are subscribed to the Google Groups >>>> "Netty discussions" group. >>>> To unsubscribe from this group and stop receiving emails from it, send an >>>> email to [email protected] >>>> <mailto:[email protected]>. >>>> To view this discussion on the web visit >>>> https://groups.google.com/d/msgid/netty/CA%2B%3DgZKAiQseAviHAkbE-eoybDWQsGY1ek6LtXkMCebcMd4WU4g%40mail.gmail.com >>>> >>>> <https://groups.google.com/d/msgid/netty/CA%2B%3DgZKAiQseAviHAkbE-eoybDWQsGY1ek6LtXkMCebcMd4WU4g%40mail.gmail.com?utm_medium=email&utm_source=footer>. >>>> For more options, visit https://groups.google.com/d/optout >>>> <https://groups.google.com/d/optout>. >>> >>> >>> -- >>> You received this message because you are subscribed to a topic in the >>> Google Groups "Netty discussions" group. >>> To unsubscribe from this topic, visit >>> https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe >>> <https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe>. >>> To unsubscribe from this group and all its topics, send an email to >>> [email protected] >>> <mailto:[email protected]>. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/netty/925935D9-92F1-4516-B9E9-524991BF053E%40googlemail.com >>> >>> <https://groups.google.com/d/msgid/netty/925935D9-92F1-4516-B9E9-524991BF053E%40googlemail.com?utm_medium=email&utm_source=footer>. >>> >>> For more options, visit https://groups.google.com/d/optout >>> <https://groups.google.com/d/optout>. >>> >>> >>> -- >>> You received this message because you are subscribed to the Google Groups >>> "Netty discussions" group. >>> To unsubscribe from this group and stop receiving emails from it, send an >>> email to [email protected] >>> <mailto:[email protected]>. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/netty/CA%2B%3DgZKDfjjLkj1yytjhrU564K3Z36SOaqfuPcER-CK3_vrK16g%40mail.gmail.com >>> >>> <https://groups.google.com/d/msgid/netty/CA%2B%3DgZKDfjjLkj1yytjhrU564K3Z36SOaqfuPcER-CK3_vrK16g%40mail.gmail.com?utm_medium=email&utm_source=footer>. >>> For more options, visit https://groups.google.com/d/optout >>> <https://groups.google.com/d/optout>. >> >> >> -- >> You received this message because you are subscribed to a topic in the >> Google Groups "Netty discussions" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe >> <https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe>. >> To unsubscribe from this group and all its topics, send an email to >> [email protected] >> <mailto:[email protected]>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/netty/A77C4A03-0A98-4104-9791-7F161A3AAD91%40googlemail.com >> >> <https://groups.google.com/d/msgid/netty/A77C4A03-0A98-4104-9791-7F161A3AAD91%40googlemail.com?utm_medium=email&utm_source=footer>. >> >> For more options, visit https://groups.google.com/d/optout >> <https://groups.google.com/d/optout>. >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Netty discussions" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected] >> <mailto:[email protected]>. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/netty/CA%2B%3DgZKCA8g-4E0AJCf%2BopC2NMufmQBGcWnMXHkwbDj-veW7x8w%40mail.gmail.com >> >> <https://groups.google.com/d/msgid/netty/CA%2B%3DgZKCA8g-4E0AJCf%2BopC2NMufmQBGcWnMXHkwbDj-veW7x8w%40mail.gmail.com?utm_medium=email&utm_source=footer>. >> For more options, visit https://groups.google.com/d/optout >> <https://groups.google.com/d/optout>. > > > -- > You received this message because you are subscribed to a topic in the Google > Groups "Netty discussions" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe > <https://groups.google.com/d/topic/netty/Ve4lnRvFXjM/unsubscribe>. > To unsubscribe from this group and all its topics, send an email to > [email protected] > <mailto:[email protected]>. > To view this discussion on the web visit > https://groups.google.com/d/msgid/netty/D30BEE2E-9119-4FD3-A144-F6B765B189E8%40googlemail.com > > <https://groups.google.com/d/msgid/netty/D30BEE2E-9119-4FD3-A144-F6B765B189E8%40googlemail.com?utm_medium=email&utm_source=footer>. > > For more options, visit https://groups.google.com/d/optout > <https://groups.google.com/d/optout>. > > > -- > You received this message because you are subscribed to the Google Groups > "Netty discussions" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected] > <mailto:[email protected]>. > To view this discussion on the web visit > https://groups.google.com/d/msgid/netty/CA%2B%3DgZKADHSVtqFoqMRak34S23OsDrZTEfADSTMpKGa3bguU-ig%40mail.gmail.com > > <https://groups.google.com/d/msgid/netty/CA%2B%3DgZKADHSVtqFoqMRak34S23OsDrZTEfADSTMpKGa3bguU-ig%40mail.gmail.com?utm_medium=email&utm_source=footer>. > For more options, visit https://groups.google.com/d/optout > <https://groups.google.com/d/optout>. -- You received this message because you are subscribed to the Google Groups "Netty discussions" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/netty/E6F9AACD-5A0B-4395-BD0B-139D2A858E33%40googlemail.com. For more options, visit https://groups.google.com/d/optout.
