Hi Maarten,
thanks a lot. You were most probably right. Just some minutes before you
answered I started alraed to fiddle around with input.fl actully
destrying it due to very limited knowledge about it. After cleaning it
up and removing the job that always let the workload_manager crash the
RB now works again, at least it started assigning jobs to sites.
Best wishes,
Crhistoph
On Mon, 30 Apr 2007 16:42:54 +0200
Maarten Litmaath <[log in to unmask]> wrote:
> Christoph Wissing wrote:
>
> > Dear *,
> >
> > on one of our LCG-RBs we have the problem that the workload_manager
> > process keeps producing segmentation faults. It might be related to
> > the problems observed last week with crashing network server, but I
> > do not know.
>
> Most probably this bug:
>
> https://savannah.cern.ch/bugs/index.php?20973
>
> It has a recipe to get you out of that mess.
>
> > Running the orkload_manager in gdb I end with:
> > Program received signal SIGSEGV, Segmentation fault.
> > [Switching to Thread 37665712 (LWP 1808)]
> > 0x08116029 in std::_Construct<classad::ExprTree*,
> > classad::ExprTree*> ()
> >
> > The RB is on the most recent update level of gLite/LCG. Three other
> > RBs at DESY with same software run fine. So it might be that a user
> > found a JDL that crashes the box. I put the backtrace for experts at
> > the very end of this mail.
> >
> > Any help is welcome!
> >
> > Best regards,
> > Christoph
> >
> > The backtrace reads as follows:
> > (gdb) backtrace
> > #0 0x08116029 in std::_Construct<classad::ExprTree*,
> > #classad::ExprTree*> () 1 0x08115d80 in
> > #std::vector<classad::ExprTree*, std::allocator<classad::ExprTree*>
> > #>::push_back () 2 0x081a43ec in classad::ExprList::GetComponents
> > #() 3 0x00469333 in edg::workload::planning::plugin::doMatch ()
> > from /opt/edg/lib/libedg_wl_classad_plugin.so
> > #4 0x081aa9f3 in classad::FunctionCall::_Evaluate ()
> > #5 0x081a54a6 in classad::ExprTree::Evaluate ()
> > #6 0x081b61b0 in classad::Operation::_Evaluate ()
> > #7 0x081a54a6 in classad::ExprTree::Evaluate ()
> > #8 0x081b6140 in classad::Operation::_Evaluate ()
> > #9 0x081a54a6 in classad::ExprTree::Evaluate ()
> > #10 0x081b6140 in classad::Operation::_Evaluate ()
> > #11 0x081a54a6 in classad::ExprTree::Evaluate ()
> > #12 0x0819da1e in classad::AttributeReference::_Evaluate ()
> > #13 0x081a54a6 in classad::ExprTree::Evaluate ()
> > #14 0x081a14e4 in classad::ClassAd::EvaluateAttr ()
> > #15 0x081a1b0d in classad::ClassAd::EvaluateAttrBool ()
> > #16 0x00ec3667 in edg::workload::common::utilities::match ()
> > from /opt/edg/lib/libedg_wl_util.so.0
> > #17 0x00ec3702 in
> > #edg::workload::common::utilities::left_matches_right ()
> > from /opt/edg/lib/libedg_wl_util.so.0
> > #18 0x00d54cad in
> > #edg::workload::planning::broker::matchmaking::matchmakerGlueIm---T
> > #ype <return> to continue, or q <return> to quit---
> > pl::checkRequirement () from
> > /opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0
> > #19 0x00d39fd8 in
> > #edg::workload::planning::broker::matchmaking::MatchMaker<edg::work
> > #load
> > #::planning::broker::matchmaking::matchmakerGlueImpl>::checkRequire
> > #ment
> > () from /opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0
> > #20 0x00d3efd1 in
> > #edg::workload::planning::broker::RBSimpleImpl::findSuitableCEs ()
> > #from /opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0 21
> > #0x00d34ec4 in
> > #edg::workload::planning::broker::ResourceBroker::findSuitableCEs ()
> > #from /opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0 22
> > #0x00d2a517 in edg::workload::planning::broker::(anonymous
> > #namespace)::f_resolve_mm () from
> > #/opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0 23 0x00d2cbd8 in
> > #edg::workload::planning::broker::Helper::resolve ()
> > from /opt/edg/lib/libedg_wl_helper_broker_prefetch.so.0
> > #24 0x001d4d41 in edg::workload::planning::helper::Helper::resolve
> > #()
> > from /opt/edg/lib/libedg_wl_helper.so.0
> > #25 0x001dfef6 in
> > #edg::workload::planning::helper::RequestStateMachine::next_step ()
> > #from /opt/edg/lib/libedg_wl_helper.so.0 26 0x001de5c0 in
> > #edg::workload::planning::helper::Request::Impl::resolve ()
> > from /opt/edg/lib/libedg_wl_helper.so.0
> > #27 0x001debee in edg::workload::planning::helper::Request::resolve
> > #()
> > from /opt/edg/lib/libedg_wl_helper.so.0
> > #28 0x00f978f1 in
> > #edg::workload::planning::manager::RequestPlanningPolicy::Plan
> > () from /opt/edg/lib/libedg_wl_manager_server.so.0
> > #29 0x00f80a61 in
> > #edg::workload::planning::manager::WMReal<edg::workload::planni---T
> > #ype <return> to continue, or q <return> to quit---
> > ng::manager::RequestPlanningPolicy,
> > edg::workload::planning::manager::JCDeliveryPolicy,
> > edg::workload::planning::manager::JCCancellingPolicy>::submit ()
> > from /opt/edg/lib/libedg_wl_manager_server.so.0
> > #30 0x00c9592d in
> > #edg::workload::planning::manager::WorkloadManager::submit ()
> > from /opt/edg/lib/libedg_wl_manager_common.so.0
> > #31 0x00f7aae1 in
> > #edg::workload::planning::manager::RequestHandler::Impl::run
> > () from /opt/edg/lib/libedg_wl_manager_server.so.0
> > #32 0x00f7c938 in
> > #edg::workload::planning::manager::RequestHandler::run ()
> > from /opt/edg/lib/libedg_wl_manager_server.so.0
> > #33 0x0810ca07 in
> > #edg::workload::common::task::ReaderFunctor<std::pair<boost::functi
> > #on0< void, boost::empty_function_policy,
> > #boost::empty_function_mixin, std::allocator<boost::function_base>
> > #>, boost::shared_ptr<classad::ClassAd> > >::operator() () 34
> > #0x0810c687 in
> > #boost::detail::function::void_function_obj_invoker0<edg::workload:
> > #:com mon::task::ReaderFunctor<std::pair<boost::function0<void,
> > #boost::empty_function_policy, boost::empty_function_mixin,
> > #std::allocator<boost::function_base> >,
> > #boost::shared_ptr<classad::ClassAd> > >, void>::invoke () 35
> > #0x08214631 in boost::thread_group::join_all () 36 0x0042fdd8 in
> > #start_thread () from /lib/tls/libpthread.so.0 37 0x007a8d2a in
> > #clone () from /lib/tls/libc.so.6
> >
>
--
+-----------------------------------+
| Christoph Wissing DESY - IT |
| E-Mail: [log in to unmask] |
| Phone: +49(0)40/8998-4122 |
+-----------------------------------+
|