Error loading module on remote SSH workers

[jebej]

On julia 0.6.2, from a Windows machine, and with SSH remote workers on CentOS, it seems that I cannot load modules remotely, due to some missing anonymous function, which makes it hard to track down the issue. I have asked about this on Discourse, but no solution has been found.

From what I understand, code loading from the master node should still work.

[vchuravy]

Sounds like a dup of #26805, which will be fixed by #26813 (I just haven't gotten around to adding Tests)

[jebej]

Is there anyway I can verify that? Note that loading the module locally prior does not help:

julia> using OrdinaryDiffEq

julia> addprocs([("admin@some.server",2)],exename="julia",dir="/home/admin",tunnel=true)
2-element Array{Int64,1}:
 2
 3

julia> @everywhere using OrdinaryDiffEq
ERROR: On worker 2:
UndefVarError: ##695#697 not defined
deserialize_datatype at .\serialize.jl:973
handle_deserialize at .\serialize.jl:677
deserialize at .\serialize.jl:637
handle_deserialize at .\serialize.jl:684
deserialize_msg at .\distributed\messages.jl:98
message_handler_loop at .\distributed\process_messages.jl:161
process_tcp_streams at .\distributed\process_messages.jl:118
#99 at .\event.jl:73
#remotecall_fetch#141(::Array{Any,1}, ::Function, ::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:354
remotecall_fetch(::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:346
#remotecall_fetch#144(::Array{Any,1}, ::Function, ::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367
remotecall_fetch(::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367
(::##11#13)() at .\distributed\macros.jl:102

...and 1 more exception(s).

Stacktrace:
 [1] sync_end() at .\task.jl:287
 [2] macro expansion at .\distributed\macros.jl:112 [inlined]
 [3] anonymous at .\<missing>:?

[vchuravy]

Can you post a minimal working example? Something I can run myself?

…

On Fri, May 4, 2018, 11:51 Jeremy ***@***.***> wrote: Is there anyway I can verify that? Note that loading the module locally prior does not help: julia> using OrdinaryDiffEq julia> ***@***.***",2)],exename="julia",dir="/home/admin",tunnel=true) 2-element Array{Int64,1}: 2 3 julia> @Everywhere using OrdinaryDiffEq ERROR: On worker 2: UndefVarError: ##695#697 not defined deserialize_datatype at .\serialize.jl:973 handle_deserialize at .\serialize.jl:677 deserialize at .\serialize.jl:637 handle_deserialize at .\serialize.jl:684 deserialize_msg at .\distributed\messages.jl:98 message_handler_loop at .\distributed\process_messages.jl:161 process_tcp_streams at .\distributed\process_messages.jl:118 #99 at .\event.jl:73 #remotecall_fetch#141(::Array{Any,1}, ::Function, ::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:354 remotecall_fetch(::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:346 #remotecall_fetch#144(::Array{Any,1}, ::Function, ::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367 remotecall_fetch(::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367 (::##11#13)() at .\distributed\macros.jl:102 ...and 1 more exception(s). Stacktrace: [1] sync_end() at .\task.jl:287 [2] macro expansion at .\distributed\macros.jl:112 [inlined] [3] anonymous at .\<missing>:? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#26975 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAI3atABvLkGcqMvI0B4BiqYvRP6Svh2ks5tvHkQgaJpZM4Tyvou> .

[vchuravy]

The other thing you can try to do is to check if the PR I mentioned fixes your issue.

…

On Fri, May 4, 2018, 12:01 Valentin Churavy ***@***.***> wrote: Can you post a minimal working example? Something I can run myself? On Fri, May 4, 2018, 11:51 Jeremy ***@***.***> wrote: > Is there anyway I can verify that? Note that loading the module locally > prior does not help: > > julia> using OrdinaryDiffEq > > julia> ***@***.***",2)],exename="julia",dir="/home/admin",tunnel=true) > 2-element Array{Int64,1}: > 2 > 3 > > julia> @Everywhere using OrdinaryDiffEq > ERROR: On worker 2: > UndefVarError: ##695#697 not defined > deserialize_datatype at .\serialize.jl:973 > handle_deserialize at .\serialize.jl:677 > deserialize at .\serialize.jl:637 > handle_deserialize at .\serialize.jl:684 > deserialize_msg at .\distributed\messages.jl:98 > message_handler_loop at .\distributed\process_messages.jl:161 > process_tcp_streams at .\distributed\process_messages.jl:118 > #99 at .\event.jl:73 > #remotecall_fetch#141(::Array{Any,1}, ::Function, ::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:354 > remotecall_fetch(::Function, ::Base.Distributed.Worker, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:346 > #remotecall_fetch#144(::Array{Any,1}, ::Function, ::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367 > remotecall_fetch(::Function, ::Int64, ::Expr, ::Vararg{Expr,N} where N) at .\distributed\remotecall.jl:367 > (::##11#13)() at .\distributed\macros.jl:102 > > ...and 1 more exception(s). > > Stacktrace: > [1] sync_end() at .\task.jl:287 > [2] macro expansion at .\distributed\macros.jl:112 [inlined] > [3] anonymous at .\<missing>:? > > — > You are receiving this because you commented. > > > Reply to this email directly, view it on GitHub > <#26975 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/AAI3atABvLkGcqMvI0B4BiqYvRP6Svh2ks5tvHkQgaJpZM4Tyvou> > . >

[jebej]

Can you post a minimal working example? Something I can run myself?

Is what I just posted not sufficient? You would need access to remote SSH workers, since it all works fine with local workers.

[jebej]

Is there any way I could try manually loading the code on the workers to see where/why it fails?

[jebej]

The other thing you can try to do is to check if the PR I mentioned fixes
your issue.

I can try this fix as soon as it makes its way in a nightly, but it does seem like it is a different problem.

[simonbyrne]

Code loading has changed significantly from 0.6. Please open a new issue if this is still a problem.