-
Notifications
You must be signed in to change notification settings - Fork 243
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Docker Build Error on folly #4
Comments
In a VM environment, problem is fixed by increasing the VM memory (8192 MB). |
Closed
facebook-github-bot
pushed a commit
that referenced
this issue
Feb 28, 2020
Summary: There is a race condition when KvStoreWrapper and ZMQ background threads are destroyed. This diff is an attempt to explicitly destroy KvStoreWrapper in TearDown Also use unique_ptr instead of shared_ptr ``` WARNING: ThreadSanitizer: data race (pid=49594) Write of size 8 at 0x7ba000000220 by thread T1: #0 close <null> (link_monitor_test+0xdb1f6a) #1 zmq::signaler_t::~signaler_t() /home/engshare/third-party2/zeromq/4.3.1/src/zeromq-4.3.1/src/signaler.cpp:114:20 (link_monitor_test+0xbd1274) Previous read of size 8 at 0x7ba000000220 by main thread: #0 epoll_ctl <null> (link_monitor_test+0xd9bdef) #1 epoll_del /home/engshare/third-party2/libevent/1.4.14b_hphp/src/libevent-1.4.14b-stable/epoll.c:485 (link_monitor_test+0xec4253) #2 folly::EventBaseEvent::eb_event_del() <null> (link_monitor_test+0xa747ef) #3 folly::EventHandler::~EventHandler() <null> (link_monitor_test+0xa74a70) #4 std::_Hashtable<int, std::pair<int const, openr::OpenrEventBase::ZmqEventHandler>, std::allocator<std::pair<int const, openr::OpenrEventBase::ZmqEventHandler> >, std::__detail::_Select1st, std::equal_to<int>, std::hash<int>, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, std::__detail::_Hashtable_traits<false, false, true> >::~_Hashtable() <null> (link_monitor_test+0x219620) #5 openr::OpenrEventBase::~OpenrEventBase() <null> (link_monitor_test+0x212d8d) #6 openr::KvStore::~KvStore() <null> (link_monitor_test+0x1bfbb7) #7 openr::KvStore::~KvStore() <null> (link_monitor_test+0x1bfbe9) #8 openr::KvStoreWrapper::~KvStoreWrapper() <null> (link_monitor_test+0x117533) #9 std::_Sp_counted_ptr_inplace<openr::KvStoreWrapper, std::allocator<openr::KvStoreWrapper>, (__gnu_cxx::_Lock_policy)2>::_M_dispose() <null> (link_monitor_test+0x116fbd) #10 LinkMonitorTestFixture::~LinkMonitorTestFixture() <null> (link_monitor_test+0xf8285) #11 LinkMonitorTestFixture_BasicOperation_Test::~LinkMonitorTestFixture_BasicOperation_Test() <null> (link_monitor_test+0xf7ee9) #12 void testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::*)(), char const*) <null> (link_monitor_test+0xc4b2ae) #13 __libc_start_main /home/engshare/third-party2/glibc/2.26/src/glibc-2.26/csu/libc-start.c:308:16 (libc.so.6+0x211a5) ``` Reviewed By: yi-xian Differential Revision: D20148667 fbshipit-source-id: 71634c08df8ebbc98a6b8c4aa3329166834453f2
facebook-github-bot
pushed a commit
that referenced
this issue
Oct 12, 2020
Summary: As titled, several different approaches have been added to address original issue of crash when eventbase is destructed. See P135204939 ``` (gdb) bt #0 0x000000000187ea8e in re2::RE2::Set::Match (this=0x2, text=..., v=0x7fffd8c0d500, error_info=0x0) at re2/set.cc:110 #1 0x000000000072d304 in openr::KeyPrefix::keyMatch (this=<optimized out>, key=...) at openr/common/Util.cpp:50 #2 0x00000000006d3649 in openr::KvStoreFilters::keyMatch (this=0x7fb22c613660, key=..., value=...) at openr/kvstore/KvStore.cpp:69 #3 0x00000000006d60e4 in openr::KvStore::mergeKeyValues (kvStore=..., keyVals=..., filters=...) at openr/kvstore/KvStore.cpp:246 #4 0x00000000006e3813 in openr::KvStoreDb::mergePublication (this=0x7fb232fbad28, rcvdPublication=..., senderId=...) at openr/kvstore/KvStore.cpp:2835 #5 0x00000000006e32d7 in openr::KvStoreDb::processThriftSuccess (this=0x7fb232fbad28, peerName=..., pub=..., timeDelta=...) at openr/kvstore/KvStore.cpp:1395 #6 0x000000000070487b in openr::KvStoreDb::requestThriftPeerSync()::$_19::operator()(openr::thrift::Publication&&) const ( this=<optimized out>, pub=...) at openr/kvstore/KvStore.cpp:1338 #7 folly::futures::detail::wrapInvoke<openr::thrift::Publication, openr::KvStoreDb::requestThriftPeerSync()::$_19>(folly::Try<openr::thrift::Publication>&&, openr::KvStoreDb::requestThriftPeerSync()::$_19&&)::{lambda()#1}::operator()() const (this=<optimized out>) at folly/futures/Future-inl.h:99 #8 folly::futures::detail::InvokeResultWrapper<void>::wrapResult<folly::futures::detail::wrapInvoke<openr::thrift::Publication, openr::KvStoreDb::requestThriftPeerSync()::$_19>(folly::Try<openr::thrift::Publication>&&, openr::KvStoreDb::requestThriftPeerSync()::$_19&&)::{lambda()#1}>(folly::futures::detail::wrapInvoke<openr::thrift::Publication, openr::KvStoreDb::requestThriftPeerSync()::$_19>(folly::Try<openr::thrift::Publication>&&, openr::KvStoreDb::requestThriftPeerSync()::$_19&&)::{lambda()#1}) (fn=...) at folly/futures/Future-inl.h:91 #9 folly::futures::detail::wrapInvoke<openr::thrift::Publication, openr::KvStoreDb::requestThriftPeerSync()::$_19>(folly::Try<openr::thrift::Publication>&&, openr::KvStoreDb::requestThriftPeerSync()::$_19&&) (t=..., f=...) at folly/futures/Future-inl.h:109 #10 folly::Future<openr::thrift::Publication>::thenValue<openr::KvStoreDb::requestThriftPeerSync()::$_19>(openr::KvStoreDb::requestThriftPeerSync()::$_19&&) &&::{lambda(folly::Executor::KeepAlive<folly::Executor>&&, folly::Try<openr::thrift::Publication>&&)#1}::operator()(folly::Executor::KeepAlive<folly::Executor>&&, folly::Try<openr::thrift::Publication>&&) (this=<optimized out>, t=...) at folly/futures/Future-inl.h:1033 #11 folly::futures::detail::CoreCallbackState<folly::Unit, folly::Future<openr::thrift::Publication>::thenValue<openr::KvStoreDb::requestThriftPeerSync()::$_19>(openr::KvStoreDb::requestThriftPeerSync()::$_19&&) &&::{lambda(folly::Executor::KeepAlive<folly::Executor>&&, folly::Try<openr::thrift::Publication>&&)#1}>::invoke<folly::Executor::KeepAlive<folly::Executor>, folly::Try<openr::thrift::Publication> >(folly::Executor::KeepAlive<folly::Executor>&&, folly::Try<openr::thrift::Publication>&&) (this=<optimized out>, args=..., args=...) at folly/futures/Future-inl.h:145 ... #19 0x00000000015d00ee in folly::detail::function::FunctionTraits<void ()>::operator()() (this=0x7fb22c63b1f0) at folly/Function.h:416 #20 folly::EventBase::FunctionLoopCallback::runLoopCallback (this=0x7fb22c63b1c0) at folly/io/async/EventBase.h:188 #21 folly::EventBase::runLoopCallbacks (this=<optimized out>) at folly/io/async/EventBase.cpp:703 #22 folly::EventBase::loopBody (this=0x7fb23ecb9410, flags=1, ignoreKeepAlive=false) at folly/io/async/EventBase.cpp:402 #23 0x00000000015cdc60 in folly::EventBase::loopOnce (this=0x7fb23ecb9410, flags=0) at folly/io/async/EventBase.cpp:330 #24 folly::EventBase::~EventBase (this=0x7fb23ecb9410, vtt=<optimized out>) at folly/io/async/EventBase.cpp:211 #25 0x00000000006f284e in openr::KvStore::~KvStore (this=0x7fb23ecb9400) at openr/kvstore/KvStore.h:532 ``` To fix this, we introduced map to hold every individual future from thrift client. However, this occasionally will make KvStore destruction stuck when waiting for all futures to be fulfilled. We should NOT track every individual future, which is NOT necessary at all. From the crash trace, clearly, we are doing `mergePublications()` when invoking `processThriftPublication()` before checking if the peer is valid or NOT. Fix: Ignore the rest logic of callback `processThriftPublication()` if peerName is NOT valid. Reviewed By: saifhhasan Differential Revision: D24262710 fbshipit-source-id: fa69aaa5c6e43cfc861de7431b9c1e26195684a0
facebook-github-bot
pushed a commit
that referenced
this issue
Jul 26, 2023
…_ destroyed Summary: # background Open/R has been haunted with unclean exit issue carried over multiple oncall rotations. Especially the notorious one in `KvStore`, like the following gdb traces happened inside T158806075 ``` [[email protected] ~/local/fbsource (470cb762a)]$ fboss-dbg-helper gdb fsw008.p062.f01.rva2 0) /home/xiangxu1121/local/fboss_dbg/fsw008.p062.f01.rva2/snapshot_2023-07-21-15:42:58 Enter the file's number: 0 0) /home/xiangxu1121/local/fboss_dbg/fsw008.p062.f01.rva2/snapshot_2023-07-21-15:42:58/cores/dogpile.unknown_tw_task.servicerouter.1583324.QUEUE_LAG.230715-025056 - Sat Jul 15 02:50:56 2023 1) /home/xiangxu1121/local/fboss_dbg/fsw008.p062.f01.rva2/snapshot_2023-07-21-15:42:58/cores/dynocat - Fri Jul 21 15:05:59 2023 2) /home/xiangxu1121/local/fboss_dbg/fsw008.p062.f01.rva2/snapshot_2023-07-21-15:42:58/cores/openr - Wed Jul 19 09:07:29 2023Enter the file's number: 2 INFO:root:Received output from file command: INFO:root:/home/xiangxu1121/local/fboss_dbg/fsw008.p062.f01.rva2/snapshot_2023-07-21-15:42:58/cores/openr: ELF 64-bit LSB core file, x86-64, version 1 (SYSV), SVR4-style, from '/usr/sbin/openr --v=1 --vmodule=BgpSerializer*=1,FiberBgp*=2 --logging=DBG1;def', real uid: 36662, effective uid: 36662, real gid: 36337, effective gid: 36337, execfn: '/usr/sbin/openr', platform: 'x86_64' (gdb) bt #0 0x00000000035aa5b4 in std::_Hashtable<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, openr::KvStoreDb<apache::thrift::Client<openr::thrift::OpenrCtrlCpp> >::KvStorePeer>, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, openr::KvStoreDb<apache::thrift::Client<openr::thrift::OpenrCtrlCpp> >::KvStorePeer> >, std::__detail::_Select1st, std::equal_to<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::hash<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::__detail::_Mod_range_hashing, std::__detail::_Default_ranged_hash, std::__detail::_Prime_rehash_policy, std::__detail::_Hashtable_traits<true, false, true> >::find(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) () at ./fbcode/third-party-buck/platform010-compat/build/libgcc/include/c++/trunk/bits/hashtable_policy.h:431 #1 0x00000000035baf56 in openr::KvStoreDb<apache::thrift::Client<openr::thrift::OpenrCtrlCpp> >::processThriftFailure(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, folly::basic_fbstring<char, std::char_traits<char>, std::allocator<char>, folly::fbstring_core<char> > const&, std::chrono::duration<long, std::ratio<1l, 1000l> >) () at ./fbcode/third-party-buck/platform010-compat/build/libgcc/include/c++/trunk/bits/unordered_map.h:869 #2 0x00000000035c4fac in openr::KvStoreDb<apache::thrift::Client<openr::thrift::OpenrCtrlCpp> >::requestThriftPeerSync()::{lambda(folly::exception_wrapper const&)#1}::operator()(folly::exception_wrapper const&) const () at fbcode/openr/kvstore/KvStore-inl.h:1753 #3 0x00000000035c4d0c in _ZZNO5folly6FutureINS_4UnitEE13thenErrorImplIZN5openr9KvStoreDbIN6apache6thrift6ClientINS4_6thrift12OpenrCtrlCppEEEE21requestThriftPeerSyncEvEUlRKNS_17exception_wrapperEE_EENSt9enable_ifIXntsr20isFutureOrSemiFutureINS_13invoke_detail6traitsIT_E6resultISD_EEEE5valueES2_E4typeEOSK_NS_7futures6detail18InlineContinuationEENUlONS_8Executor9KeepAliveISU_EEONS_3TryIS1_EEE_clESX_S10_ () at fbcode/folly/futures/Future-inl.h:137 #4 0x00000000035c4bea in _ZN5folly6detail8function14FunctionTraitsIFvRNS_7futures6detail8CoreBaseEONS_8Executor9KeepAliveIS7_EEPNS_17exception_wrapperEEE7callBigIZNS4_4CoreINS_4UnitEE11setCallbackIZNOS_6FutureISH_E13thenErrorImplIZN5openr9KvStoreDbIN6apache6thrift6ClientINSN_6thrift12OpenrCtrlCppEEEE21requestThriftPeerSyncEvEUlRKSB_E_EENSt9enable_ifIXntsr20isFutureOrSemiFutureINS_13invoke_detail6traitsIT_E6resultISB_EEEE5valueESL_E4typeEOS12_NS4_18InlineContinuationEEUlSA_ONS_3TryISH_EEE_EEvS18_OSt10shared_ptrINS_14RequestContextEES19_EUlS6_SA_SC_E_EEvS6_SA_SC_RNS1_4DataE () at fbcode/folly/futures/detail/Core.h:619 warning: Could not find DWO CU buck-out/v2/gen/fbcode/ad98fc21e927c889/folly/futures/detail/__core__/__objects__/Core.cpp.o(0xb8563f2cc6214b1a) referenced by CU at offset 0x10ac73 [in module /data/users/xiangxu1121/fboss_dbg/packages/openr:325/openr.debuginfo] #5 0x0000000002408dc9 in folly::futures::detail::CoreBase::doCallback(folly::Executor::KeepAlive<folly::Executor>&&, folly::futures::detail::State)::$_0::operator()(folly::Executor::KeepAlive<folly::Executor>&&) () at fbcode/folly/Function.h:375 warning: Could not find DWO CU buck-out/v2/gen/fbcode/ad98fc21e927c889/folly/io/async/__async_base__/__objects__/EventBase.cpp.o(0x7990ef623e52da16) referenced by CU at offset 0x1029d6 [in module /data/users/xiangxu1121/fboss_dbg/packages/openr:325/openr.debuginfo] #6 0x000000000266571d in folly::EventBase::loopMain(int, bool) () at fbcode/folly/Function.h:375 #7 0x0000000002d20cc3 in folly::EventBase::loopOnce(int) () at fbcode/folly/io/async/EventBase.cpp:345 #8 0x0000000002542c12 in folly::EventBase::~EventBase() () at fbcode/folly/io/async/EventBase.cpp:195 #9 0x00000000035a420e in openr::KvStore<apache::thrift::Client<openr::thrift::OpenrCtrlCpp> >::~KvStore() () at fbcode/openr/kvstore/KvStore.h:629 #10 0x0000000003585a9e in main () at ./fbcode/third-party-buck/platform010-compat/build/libgcc/include/c++/trunk/bits/unique_ptr.h:85 ``` # RCA High level speaking, the issue happens with the following order: - Step1: `KvStoreDb::stop()` destructed `thriftPeers_` obj - Step2: `KvStore::stop()` invokes `OpenrEventBase::stop()`, which destructs `folly::EventBase`, which will call `loopOnce` to clean up injected callbacks in the eventbase. - Step3: In crash scenario, there are >1 callbacks being invoked(e.g. `processThriftSuccess` or `processThriftFailure`). - Step4: Sample failures are: - `processThriftSuccess` P576623566 - `processThriftFailure` P592005831 We can see there are callbacks being invoked with: - `requestThriftPeerSync` - `keepAlive` # Fix The fix should be straight-forward, which is to cancel the scheduled cb from `folly::AsyncTimeout` perspective via `reset()` or `cancelTimeout()` call. NOTE: for the keepAlive `AsyncTimeout`, we can consider remove it by leveraging the socket keepalive option and simplify this logic. Reviewed By: TangoRoxy Differential Revision: D47691545 fbshipit-source-id: 67108399a7ce9f96d08682d45d03799e6ef608fd
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Resolution: Increase memory.
Omitting the majority of the log file, will post it if wanted. Build appears to be failing in folly directory.
Built in brand new ubuntu/xenial64 Vagrant environment, with Docker installed.
Command adapted from fbcode_builder instructions. Used appropriate variables from travis.yml due to Travis script not working out of box. #3
Omitting output.
Any ideas?
The text was updated successfully, but these errors were encountered: