@Guangya, I tried with default_accepted_resource_roles option set to `dev`; same result, marathon can not start new tasks, waiting for resources.
@Guangya, @Haosdent, yes, on slave side, slave reports the resource consumed comes from `*` role; ``` [root@localhost ~]# curl http://10.132.46.71:5050/slaves | python -m json.tool % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 1091 100 1091 0 0 179k 0 --:--:-- --:--:-- --:--:-- 213k { "slaves": [ { "active": true, "attributes": {}, "hostname": "10.132.46.71", "id": "439a851d-aa09-47af-bc68-d5588dc9f5da-S0", "offered_resources": { "cpus": 0.0, "disk": 0.0, "mem": 0.0 }, "offered_resources_full": [], "pid": "slave(1)@10.132.46.71:5051", "registered_time": 1462091019.95277, "reregistered_time": 1462091019.95438, "reserved_resources": { "dev": { "cpus": 2.0, "disk": 0.0, "mem": 1024.0 } }, "reserved_resources_full": { "dev": [ { "name": "cpus", "reservation": {}, "role": "dev", "scalar": { "value": 2.0 }, "type": "SCALAR" }, { "name": "mem", "reservation": {}, "role": "dev", "scalar": { "value": 1024.0 }, "type": "SCALAR" } ] }, "resources": { "cpus": 4.0, "disk": 46055.0, "mem": 6727.0, "ports": "[31000-32000]" }, "unreserved_resources": { "cpus": 2.0, "disk": 46055.0, "mem": 5703.0, "ports": "[31000-32000]" }, "used_resources": { "cpus": 1.0, "disk": 0.0, "mem": 128.0, "ports": "[31994-31994]" }, "used_resources_full": [ { "name": "cpus", "role": "*", "scalar": { "value": 1.0 }, "type": "SCALAR" }, { "name": "mem", "role": "*", "scalar": { "value": 128.0 }, "type": "SCALAR" }, { "name": "ports", "ranges": { "range": [ { "begin": 31994, "end": 31994 } ] }, "role": "*", "type": "RANGES" } ], "version": "0.28.1" } ] } ``` while, on the master side, when curl http://10.132.46.71:5050/roles, it shows the resource comes from `dev` role; ``` [root@localhost ~]# curl http://10.132.46.71:5050/roles | python -m json.tool % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 250 100 250 0 0 46772 0 --:--:-- --:--:-- --:--:-- 50000 { "roles": [ { "frameworks": [], "name": "*", "resources": { "cpus": 0, "disk": 0, "mem": 0 }, "weight": 1.0 }, { "frameworks": [ "439a851d-aa09-47af-bc68-d5588dc9f5da-0002" ], "name": "dev", "resources": { "cpus": 1.0, "disk": 0, "mem": 128.0, "ports": "[31994-31994]" }, "weight": 1.0 } ] } ``` moreover, when i reserve all resources(dynamic reservation) from default role to `dev` role, not a single task can start on that slave, either declaring use `dev` resource or `*` resource. ``` "reserved_resources": { "dev": { "cpus": 4.0, "disk": 0.0, "mem": 6727.0 } }, "reserved_resources_full": { "dev": [ { "name": "mem", "reservation": {}, "role": "dev", "scalar": { "value": 6727.0 }, "type": "SCALAR" }, { "name": "cpus", "reservation": {}, "role": "dev", "scalar": { "value": 4.0 }, "type": "SCALAR" } ] }, "resources": { "cpus": 4.0, "disk": 46055.0, "mem": 6727.0, "ports": "[31000-32000]" }, "unreserved_resources": { "cpus": 0.0, "disk": 46055.0, "mem": 0.0, "ports": "[31000-32000]" }, ``` @Haosdent, can u post your master:5050/slaves? i think this should be the issue. -- Best Regards, GaoJin CAO ChinaMobile(Suzhou) R&D Mail：caogaojin at cmss.chinamobile.com <http://cmss.chinamobile.com/> > On May 1, 2016, at 15:32, Guangya Liu <[email protected]> wrote: > > Not sure if you need add a new parameter default_accepted_resource_roles , > even if you start marathon with role dev, it can still get resources from > both role dev and *. > > Please refer to > https://github.com/mesosphere/marathon/blob/master/docs/docs/command-line-flags.md > > <https://github.com/mesosphere/marathon/blob/master/docs/docs/command-line-flags.md> > for marathon configuration. > > v0.9.0 --default_accepted_resource_roles (Optional. Default: all roles): > Default for the "acceptedResourceRoles"attribute as a comma-separated list of > strings. All app definitions which do not specify this attribute explicitly > use this value for launching new tasks. Examples: *, production,*, production > > On Sun, May 1, 2016 at 3:21 PM, haosdent <[email protected] > <mailto:[email protected]>> wrote: > could you show the result of curl 10.132.46.71:5050/ > <http://10.132.46.71:5050/slaves>roles It shows consume "dev" in my side. > > On Sun, May 1, 2016 at 3:04 PM, caogaojin <[email protected] > <mailto:[email protected]>> wrote: > @Haosdent, by the way, i’m using mesos:0.28.1, not the master branch. and > marathon version is 1.1.1 > > -- > Best Regards, > > GaoJin CAO > ChinaMobile(Suzhou) R&D > Mail：caogaojin at cmss.chinamobile.com <http://cmss.chinamobile.com/> >> On May 1, 2016, at 14:04, haosdent <[email protected] >> <mailto:[email protected]>> wrote: >> >> Hi, @caogaoin. I try in with the master branch of Mesos and Marathon. And it >> works fine for me. >> >> >> ``` >> INFO Finished processing dda04940-55b6-4ba7-a7bb-75c2108f3151-O8. Matched 1 >> ops after 2 passes. cpus(dev) 2.0; mem(dev) 1024.0; cpus(*) 5.0; mem(*) >> 29808.0; disk(*) 915273.0; ports(*) 31000->31254,31256->32000 left. >> (mesosphere.marathon.core.matcher.manager.impl.OfferMatcherManagerActor:marathon-akka.actor.default-dispatcher-2) >> ``` >> >> Let me list my verify steps to help you find your problem when start your >> tasks. >> >> 1. start Mesos master and agent, and reserve resources for dev. >> >> ``` >> $ ./bin/mesos-master.sh --work_dir=/tmp/mesos --ip=127.0.0.1 >> --hostname=127.0.0.1 >> >> $ ./bin/mesos-slave.sh --work_dir=/tmp/mesos --ip=127.0.0.1 >> --hostname=127.0.0.1 --master=127.0.0.1:5050 <http://127.0.0.1:5050/> >> >> >> $ curl -i -d slaveId=dda04940-55b6-4ba7-a7bb-75c2108f3151-S0 \ >> -d resources='[ >> { >> "name": "cpus", >> "type": "SCALAR", >> "scalar": { >> "value": 2.0 >> }, >> "role": "dev", >> "reservation": { >> } >> >> }, >> { >> "name": "mem", >> "type": "SCALAR", >> "scalar": { >> "value": 1024.0 >> >> }, >> "role": "dev", >> "reservation": { >> } >> } >> ]' -X POST http://127.0.0.1:5050/reserve <http://127.0.0.1:5050/reserve> >> >> 2. start Marathon with dev role >> >> ``` >> ./bin/start --master 127.0.0.1:5050 <http://127.0.0.1:5050/> --zk >> zk://10.10.120.61:2181/marathon <http://10.10.120.61:2181/marathon> >> --mesos_role dev >> ``` >> >> 3. start task >> >> ``` >> $ cat task.json >> { >> "id": "/test-health", >> "cmd": "sleep 200", >> "cpus": 1.0, >> "mem": 128.0 >> } >> >> $ curl -X POST -H "Content-type: application/json" >> http://127.0.0.1:8080/v2/apps <http://127.0.0.1:8080/v2/apps> -d @task.json >> ``` >> >> On Sun, May 1, 2016 at 12:18 PM, haosdent <[email protected] >> <mailto:[email protected]>> wrote: >> Sorry for make you confusing before. The offer looks match resource >> requirement according to your log, let me reproduce it first. >> >> On Sun, May 1, 2016 at 8:57 AM, Guangya Liu <[email protected] >> <mailto:[email protected]>> wrote: >> Gaojin, >> >> Some questions and suggestions here: >> >> 1) How did you deploy the nginx with new marathon URI? I did not found that >> I can configure the new marathon URI with nginx. >> >> 2) Can you please send out the curl output for your master node: >> a) curl "http://master-ip:5050/master/slaves >> <http://9.21.51.123:5050/master/slaves>" 2>/dev/null|python -m json.tool >> >> b) curl "http://master-ip:5050/master/state >> <http://9.21.51.123:5050/master/slaves>" 2>/dev/null|python -m json.tool >> >> 3) For open source DC/OS issues, it is suggested to post your question to >> [email protected] <mailto:[email protected]> where you can always get instant and >> professional support. >> >> Thanks, >> >> Guangya >> >> >> On Sun, May 1, 2016 at 1:49 AM, caogaojin <[email protected] >> <mailto:[email protected]>> wrote: >> Hi list, >> >> I’m currently using dynamic reservation mechanism with mesosphere/DCOS 1.7, >> the opensourced version, >> and something wired just happened. I tried to setup a new marathon service >> from the universe repo, >> and i did a slight modification, setting mesos_role to `dev`. i also turned >> off mesos anthentication >> mechenism. >> >> My steps are: >> >> First i tried to reserve 2 cpu and 1743 ram to `dev` role using http >> endpoint, and it works: >> # curl -i -d slaveId=e6d590b1-5f2d-4be9-93fd-2fac364584cf-S11 \ >> -d resources='[ >> { >> "name": "cpus", >> "type": "SCALAR", >> "scalar": { >> "value": 2.0 >> }, >> "role": "dev", >> "reservation": { >> } >> >> }, >> { >> "name": "mem", >> "type": "SCALAR", >> "scalar": { >> "value": 1024.0 >> >> }, >> "role": "dev", >> "reservation": { >> } >> } >> ]' -X POST http://10.132.46.81:5050/unreserve >> <http://10.132.46.81:5050/unreserve> >> >> Second, the slave shows following infomation: >> # curl http://10.132.46.81/mesos/slaves <http://10.132.46.81/mesos/slaves> >> reserved_resources_full: { >> dev: [ >> { >> name: "cpus", >> type: "SCALAR", >> scalar: { >> value: 2 >> }, >> role: "dev", >> reservation: { } >> }, >> { >> name: "mem", >> type: "SCALAR", >> scalar: { >> value: 1743 >> }, >> role: "dev", >> reservation: { } >> } >> ] >> }, >> >> Third, i tried to run nginx using the new marathon service, which i also set >> the task only accept `dev` resources, >> >> ID: /nginx >> ... >> Resource Roles: dev >> >> 4. on marathon ui i just see the waiting status, which means there’s no >> suitable resource offer, marathon reports: >> >> Insufficient resources for [/nginx] (need cpus=1.0, mem=128.0, disk=0.0, >> ports=(1 dynamic), available in offer: [id { value: >> "e6d590b1-5f2d-4be9-93fd-2fac364584cf-O7476" } framework_id { value: >> "e6d590b1-5f2d-4be9-93fd-2fac364584cf-0006" } slave_id { value: >> "e6d590b1-5f2d-4be9-93fd-2fac364584cf-S11" } hostname: "10.132.46.84" >> resources { name: "ports" type: RANGES ranges { range { begin: 8182 end: >> 32000 } } role: "dev" reservation { } } resources { name: "cpus" type: >> SCALAR scalar { value: 2.0 } role: "dev" reservation { } } resources { name: >> "mem" type: SCALAR scalar { value: 1743.0 } role: "dev" reservation { } } >> resources { name: "ports" type: RANGES ranges { range { begin: 1025 end: >> 2180 } range { begin: 2182 end: 3887 } range { begin: 3889 end: 5049 } range >> { begin: 5052 end: 6326 } range { begin: 6328 end: 8079 } range { begin: >> 8082 end: 8180 } } role: "*" } resources { name: "disk" type: SCALAR scalar >> { value: 41497.0 } role: "*" } resources { name: "cpus" type: SCALAR scalar >> { value: 1.0 } role: "*" } resources { name: "mem" type: SCALAR scalar { >> value: 896.0 } role: "*" } url { scheme: "http" address { hostname: >> "10.132.46.84" ip: "10.132.46.84" port: 5051 } path: "/slave(1)" }] >> (mesosphere.mesos.TaskBuilder:marathon-akka.actor.default-dispatcher-19) >> >> can anyone help me with this? >> >> thank you. >> >> >> GaoJin CAO >> ChinaMobile(Suzhou) R&D >> Phone：18896723732 >> Mail：caogaojin at cmss.chinamobile.com <http://cmss.chinamobile.com/> >> >> >> >> >> -- >> Best Regards, >> Haosdent Huang >> >> >> >> -- >> Best Regards, >> Haosdent Huang > > > > > -- > Best Regards, > Haosdent Huang >
