Sepia » History » Revision 87
Revision 86 (Cron Tab, 05/18/2016 11:51 PM) → Revision 87/220 (Cron Tab, 05/19/2016 01:00 PM)
h3. Crontab<pre></pre> Crontab<pre>+---------------------------------------------------+------------------------------------------------------------------------------------------------+ | path | desc | +---------------------------------------------------+------------------------------------------------------------------------------------------------+ | ├── parallel | | | │ ├── % | | | │ ├── 0-cluster | | | │ │ ├── + | | | │ │ ├── openstack.yaml | | | │ │ └── start.yaml | Run ceph on two nodes, | | | with a separate client 0,1,2 third node. | | | Use xfs beneath the osds. | | | CephFS tests running on client 2,3 | | | | | │ ├── 1-infernalis-install | | | │ │ └── infernalis.yaml | install ceph/infernalis latest | | | run workload and upgrade-sequence in parallel | | | upgrade the client node | | | | | │ ├── 2-workload | | | │ │ ├── + | | | │ │ ├── blogbench.yaml | run a cephfs stress test | | | mount ceph-fuse on client.2 before running workunit | | | | | │ │ ├── ec-rados-default.yaml | run run randomized correctness test for rados operations | | | on an erasure-coded pool | | | | | │ │ ├── rados_api.yaml | object class functional tests | | | | | │ │ ├── rados_loadgenbig.yaml | generate read/write load with rados objects ranging from 1MB to 25MB | | | | | │ │ ├── test_rbd_api.yaml | librbd C and C++ api tests | | | | | │ │ └── test_rbd_python.yaml | librbd python api tests | | | | | │ ├── 3-upgrade-sequence | | | │ │ ├── upgrade-all.yaml | upgrade the ceph cluster | | | | | │ │ └── upgrade-mon-osd-mds.yaml | upgrade the ceph cluster, | | | upgrate in two steps | | | step one ordering: mon.a, osd.0, osd.1, mds.a | | | step two ordering: mon.b, mon.c, osd.2, osd.3 | | | ceph expected to be healthy state after each step | | | | | │ ├── 4-jewel.yaml | | | │ ├── 5-final-workload | | | │ │ ├── + | | | │ │ ├── blogbench.yaml | run a cephfs stress test | | | mount ceph-fuse on client.3 before running workunit | | | | | │ │ ├── rados-snaps-few-objects.yaml | randomized correctness test for rados operations on a replicated pool with snapshots | | | | | │ │ ├── rados_loadgenmix.yaml | generate read/write load with rados objects ranging from 1 byte to 1MB | | | | | │ │ ├── rados_mon_thrash.yaml | librados C and C++ api tests | | | | | │ │ ├── rbd_cls.yaml | rbd object class functional tests | | | | | │ │ ├── rbd_import_export.yaml | run basic import/export cli tests for rbd | | | | | │ │ └── rgw_swift.yaml | swift api tests for rgw | | | | | │ └── distros | | | │ ├── centos_7.2.yaml | | | │ └── ubuntu_14.04.yaml | | | ├── point-to-point-x | | | │ ├── % | | | │ ├── distros | | | │ │ ├── centos_7.2.yaml | | | │ │ └── ubuntu_14.04.yaml | | | │ └── point-to-point.yaml | Run ceph on two nodes, using one of them as a client, | | | with a separate client-only node. | | | Use xfs beneath the osds. | | | install ceph/infernalis v9.2.0 point version | | | run workload and upgrade-sequence in parallel | | | install ceph/infernalis latest version | | | run workload and upgrade-sequence in parallel | | | install ceph/-x version (jewel) | | | run workload and upgrade-sequence in parallel | | | | | ├── stress-split | | | │ ├── % | | | │ ├── 0-cluster | | | │ │ ├── + | | | │ │ ├── openstack.yaml | | | │ │ └── start.yaml | Run ceph on two nodes, | | | with a separate client-only node. | | | Use xfs beneath the osds. | | | | | │ ├── 1-infernalis-install | | | │ │ └── infernalis.yaml | install ceph/infernalis latest | | │ ├── 2-partial-upgrade | | | │ │ └── firsthalf.yaml | install upgrade ceph/-x on one node only | | | 1st half | | | restart : osd.0,1,2,3,4,5 | | | | | │ ├── 3-thrash | | | │ │ └── default.yaml | randomly kill and revive osd | | | small chance to increase the number of pgs | | | | | │ ├── 4-mon | | | │ │ └── mona.yaml | restart mon.a so it is upgraded to -x | | | | | │ ├── 5-workload | | | │ │ ├── + | | | │ │ ├── rbd-cls.yaml | run basic cls tests for rbd | | | | | │ │ ├── rbd-import-export.yaml | run basic import/export cli tests for rbd | | | | | │ │ ├── readwrite.yaml | randomized correctness test for rados operations on a replicated pool, | | | using only reads, writes, and deletes | | | | | │ │ └── snaps-few-objects.yaml | randomized correctness test for rados operations on a replicated pool with snapshot operations | | | | | │ ├── 6-next-mon | | | │ │ └── monb.yaml | restart mon.b so it is upgraded to -x | | | | | │ ├── 7-workload | | | │ │ ├── + | | | │ │ ├── radosbench.yaml | run randomized correctness test for rados operations | | | generate write load with rados bench | | | | | │ │ └── rbd_api.yaml | librbd C and C++ api tests | | | | | │ ├── 8-next-mon | | | │ │ └── monc.yaml | restart mon.c so it is upgraded to -x | | | as all mon were upgrated, expected ceph cluster reach quorum | | | | | │ ├── 9-workload | | | │ │ ├── + | | | │ │ ├── rbd-python.yaml | librbd python api tests | | | | | │ │ ├── rgw-swift.yaml | swift api tests for rgw | | | | | │ │ └── snaps-many-objects.yaml | randomized correctness test for rados operations on a replicated pool with snapshot operations | | | | | │ └── distros | | | │ ├── centos_7.2.yaml | | | │ └── ubuntu_14.04.yaml | | | ├── stress-split-erasure-code | | | │ ├── % | | | │ ├── 0-cluster | | | │ │ ├── + | | | │ │ ├── openstack.yaml | | | │ │ └── start.yaml | Run ceph on two nodes, | | | with a separate client-only node. | | | Use xfs beneath the osds. | | | | | │ ├── 1-infernalis-install | | | │ │ └── infernalis.yaml | install ceph/infernalis latest | | │ ├── 2-partial-upgrade | | | │ │ └── firsthalf.yaml | install upgrade ceph/-x on one node only | | | 1st half | | | restart : osd.0,1,2,3,4,5 | | | | | │ ├── 3-thrash | | | │ │ └── default.yaml | randomly kill and revive osd | | | small chance to increase the number of pgs | | | | | │ ├── 4-mon | | | │ │ └── mona.yaml | restart mon.a so it is upgraded to -x | | | | | │ ├── 5-workload | | | │ │ └── ec-rados-default.yaml | randomized correctness test for rados operations on an erasure coded pool | | | | | │ ├── 6-next-mon | | | │ │ └── monb.yaml | restart mon.b so it is upgraded to -x | | | | | │ ├── 8-next-mon | | | │ │ └── monc.yaml | restart mon.c so it is upgraded to -x | | | as all mon were upgrated, expected ceph cluster reach quorum | | | | | │ ├── 9-workload | | | │ │ └── ec-rados-plugin=jerasure-k=3-m=1.yaml | randomized correctness test for rados operations on an erasure coded pool | | | using the jerasure plugin with k=3 and m=1 | | | | | │ └── distros | | | │ ├── centos_7.2.yaml | | | │ └── ubuntu_14.04.yaml | | | └── stress-split-erasure-code-x86_64 | | | ├── % | | | ├── 0-cluster | | | │ ├── + | | | │ ├── openstack.yaml | | | │ └── start.yaml | Run ceph on two nodes, | | | with a separate client-only node. | | | Use xfs beneath the osds. | | | | | ├── 0-x86_64.yaml | | | ├── 1-infernalis-install | | | │ └── infernalis.yaml | install ceph/infernalis latest | | ├── 2-partial-upgrade | | | │ └── firsthalf.yaml | install upgrade ceph/-x on one node only | | | 1st half | | | restart : osd.0,1,2,3,4,5 | | | | | ├── 3-thrash | | | │ └── default.yaml | randomly kill and revive osd | | | small chance of increasing the number of pgs | | | | | ├── 4-mon | | | │ └── mona.yaml | restart mon.a so it is upgraded to -x | | | | | ├── 5-workload | | | │ └── ec-rados-default.yaml | randomized correctness test for rados operations on an erasure coded pool | | | | | ├── 6-next-mon | | | │ └── monb.yaml | restart mon.b so it is upgraded to -x | | | | | ├── 8-next-mon | | | │ └── monc.yaml | restart mon.c so it is upgraded to -x | | | as all mon were upgrated, expected ceph cluster reach quorum | | | | | └── 9-workload | | | └── ec-rados-plugin=jerasure-k=3-m=1.yaml | randomized correctness test for rados operations on an erasure coded pool | | | using the jerasure plugin with k=3 and m=1 | | | | +---------------------------------------------------+------------------------------------------------------------------------------------------------+</pre>