You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Scott Smith <sc...@gmail.com> on 2012/04/21 02:18:32 UTC

slave registering itself multiple times

Running mesos #1205738 (should I upgrade? if so, will it still work
with the git head of Spark?)

I find often mesos-slave's are getting registered multiple times:

I0420 23:52:37.429045 20020 master.cpp:814] Attempting to register
slave 201204202347-0-8 at slave@10.252.94.24:50961
I0420 23:52:37.429446 20020 master.cpp:1057] Master now considering a
slave at ip-10-252-94-24.us-west-2.compute.internal:50961 as active
I0420 23:52:37.429523 20020 master.cpp:1588] Adding slave
201204202347-0-8 at ip-10-252-94-24.us-west-2.compute.internal with
cpus=2; mem=1184
I0420 23:52:37.429679 20020 simple_allocator.cpp:71] Added slave
201204202347-0-8 with cpus=2; mem=1184
I0420 23:52:38.429553 20020 master.cpp:814] Attempting to register
slave 201204202347-0-9 at slave@10.252.94.24:50961
I0420 23:52:38.429792 20020 master.cpp:1588] Adding slave
201204202347-0-9 at ip-10-252-94-24.us-west-2.compute.internal with
cpus=2; mem=1184
I0420 23:52:38.429898 20020 simple_allocator.cpp:71] Added slave
201204202347-0-9 with cpus=2; mem=1184
I0420 23:52:39.430135 20020 master.cpp:814] Attempting to register
slave 201204202347-0-10 at slave@10.252.94.24:50961
I0420 23:52:39.430416 20020 master.cpp:1588] Adding slave
201204202347-0-10 at ip-10-252-94-24.us-west-2.compute.internal with
cpus=2; mem=1184
I0420 23:52:39.430531 20020 simple_allocator.cpp:71] Added slave
201204202347-0-10 with cpus=2; mem=1184
I0420 23:52:40.431375 20020 master.cpp:814] Attempting to register
slave 201204202347-0-11 at slave@10.252.94.24:50961
I0420 23:52:40.431562 20020 master.cpp:1588] Adding slave
201204202347-0-11 at ip-10-252-94-24.us-west-2.compute.internal with
cpus=2; mem=1184
I0420 23:52:40.431658 20020 simple_allocator.cpp:71] Added slave
201204202347-0-11 with cpus=2; mem=1184

I know for a fact mesos-slave is only being run once; I added a 'date
>> file' line to the script that launches it and checked the file.  It
doesn't matter if I add a sleep, nohup, and/or stdin/out/err
redirections.

I have attached the tcpdump from that session.
-- 
        Scott