Operate Apps¶
Selecting Objects¶
All¶
om '**' ls
All Services¶
om '*' ls
Single Service¶
om <path> ls
om ns1/svc/web1 ls
Object list¶
om <path1>,<path2> ls
om ns1/svc/web1,ns1/vol/wev1 ls
Services by State¶
om '*' ls --status down
List all services in 'down' state.
om '*' ls --status up,warn
List all services in 'up' and 'warn' state.
Service Selector Expressions¶
om <expr> ls
Where <expr> is:
<path glob pattern>[!]<param><op><value>
Where <param> is:
<rid>.<key>in a service config file<group>.<key>in a service config file, where<group>is a driver group name like disk, fs, task, ... or an empty string, meaning "all resources"<key>in the service config file header
Where <op> is:
<><=>==operators:as a existance test operator (empty value)~as a regexp operator
And:
!as a negation operator+as the AND expression separator,as the OR expression separator
The matching is case-sensitive except for booleans.
Examples:
services with name ending with dns or starting with ha and with an app resource with a timeout set superior to 1:
$ om '*dns,ha*+app.timeout>1' ls
ha1
ha2
ha3
pridns
Services with at least one ip resource and one task resource:
$ om 'ip:+task:' ls
ha1
ha2
ha3
registry
Services with at least one monitored resource and monitor_schedule not set:
$ om '!monitor_schedule+.monitor=true' ls
ha1
ha4
Services Status¶
Cluster Overview¶
Human Readable¶
om mon
This command fetches the information from the daemon listener. Hence,
- The daemon must be up and running
- The displayed information is near synchronous
Threads aubergine nuc hb#1.rx running 224.3.29.71:10001 | / X hb#1.tx running 224.3.29.71:10001 | / O hb#2.rx running 0.0.0.0:10004 | / O hb#2.tx running | / O listener running 0.0.0.0:1214 monitor running scheduler running Nodes aubergine nuc 15m | 0.6 0.0 state | Services aubergine nuc collector up failover | O^ ha1 warn!^ failover | !! !!^ pridns up failover | O^ registry up failover | O^ testapplim n/a flex | /^ testapplim2 n/a flex | /* /* testbnp n/a failover | /P testdrbd n/a failover | /P /P testmd up! flex | X! O!^ testmd2 up!^ failover | O! X!^
To optimize information density, attributes, status and alerts are represented as markers.
Heartbeats markers:
| Marker | On hb.tx target | On hb.rx source |
|---|---|---|
O |
data has been sent in time | data has been received in time |
X |
data has not been sent in time | data has not been received in time |
/ |
not applicable | not applicable |
Nodes, services and instances markers:
| Marker | On service instance | On Service | On node status |
|---|---|---|---|
O |
up | ||
o |
standby up instance | ||
X |
down instance or heartbeat | ||
x |
standby down instance | ||
/ |
not applicable, undefined | ||
^ |
placement leader | placement alert | |
! |
warning | warning raise by any instance | |
!! |
not fully available instance | ||
* |
frozen instance | frozen node | |
P |
not fully provisioned instance |
Machine readable¶
om daemon status --format json
om daemon status --format flat_json
Watch¶
om mon --watch
Stats¶
om mon --stats
Actions¶
Base Actions¶
Start¶
om <path> start --local
Start the local service instance, shortcutting the orchestrator. Resources start order is ip, disk, fs, share, container, app.
om <path> start [--wait] [--time <duration expr>] [--watch]
Tell the orchestrator to start the service on the nodes the placement policy and constraints choose.
By default, the command returns as soon has to daemon has acknowedged the order. With --wait, the command will wait for the action completion before returning. --time set a maximum wait time.
Stop¶
om <path> stop --local
Stop the local service instance, shortcutting the orchestrator. Resources stop order is app, container, share, fs, disk, ip.
om <path> stop [--wait] [--time <duration expr>] [--watch]
Tell the orchestrator to stop the service wherever it runs and freeze it so it is not restarted.
By default, the command returns as soon has to daemon has acknowedged the order. With --wait, the command will wait for the action completion before returning. --time set a maximum wait time.
Relocation¶
om <path> switch --node <nodename> [--wait] [--time <duration expr>] [--watch]
Stop the service on <nodename> peers and start it on <nodename>.
All service instances are thawed at the end of this ended, whatever their initial frozen state.
om <path> takeover [--wait] [--time <duration expr>] [--watch]
Stop the service instances on peers and start it on the local node.
All service instances are thawed at the end of this ended, whatever their initial frozen state.
om <path> giveback [--wait] [--time <duration expr>] [--watch]
Thaw the nodes and service instances, stop the service instances running on non-leader nodes, and let the orchestrator start the instances on the leaders.
All service instances are thawed at the end of this ended, whatever their initial frozen state.
Handling Failures¶
When an action is submitted to the agent daemons, they orchestrate the execution plan to make the service reach the desired state. If a step of this plan fails, the orchestrator is blocked, the failure reported in om mon and om <path> print status, and the target state is still set.
For example, the svc1 failover service is requested to start. The n1 node is the leader and its instance started, but the action fails. This instance service monitor status transitioned to start failed, and the orchestration is blocked.
To let the daemon retry the execution plan, the failure can be cleared, using:
om <path> clear
To abort the action, use:
om <path> abort
Sync¶
om <path> sync all
Run the sync resources replication to all targets, either prd or drp.
om <path> sync nodes
Trigger hard-coded and user-defined file synchronization to secondary nodes. Optionally creates snapshots to send a coherent file set. No-op if run from a node not running the service.
om <path> sync drp
Trigger hard-coded and user-defined file synchronization to disaster recovery nodes. Optionally creates snapshots to send a coherent file set. No-op if run from a node not running the service.
See also
Resource Filtering¶
om <path> --rid <rid>[,<rid>,...] <action>
Execute <action> on <path> resources specified by --rid.
om <path> --rid <drvgrp>[,<drvgrp>,...] <action>
Execute <action> on <path> resources of driver groups specified by by --rid.
The supported driver groups are:
- ip
- disk
- fs
- share
- container
- app
- sync
- task
Resource identifiers and driver groups can be mixed in a --rid expression.
om <path> --tags tag1,tag2 <action>
Execute <action> on <path> resources tagged with either tag1 or tag2.
om <path> --tags tag1+tag2,tag3 <action>
Execute <action> on <path> resources tagged with both tag1 or tag2 or with tag3.
om <path> --subsets s1,s2 <action>
Execute <action> on <path> resources in subset s1 or s2
Logging¶
All action logs are multiplexed to:
- stdout/stderr
<OSVCLOG>/<name>.logfor svc kind in root namespace.<OSVCLOG>/<kind>/<name>.logfor non-svc kind in root namespace.<OSVCLOG>/namespaces/<namespace>/<kind>/<name>.logfor namespaced objects. Daily rotation on these files, and size limit rotation- collector database Optional, through asynchronous xmlrpc calls.
- syslog
Optional, disabled by default, configured in
node.conf
Examples¶
Print resource status of a service:
$ om osvprdcollector.opensvc.com print status
osvprdcollector.opensvc.com
overall up
|- avail up
| |- ip#0 .... up 37.59.71.25@br0@container#0
| |- fs#1 .... stdby up zfs data/osvprdcollector.opensvc.com@/srv/osvprdcollector.opensvc.com
| |- fs#3 .... stdby up zfs data/osvprdcollector.opensvc.com/data@/srv/osvprdcollector.opensvc.com/data
| |- fs#2 .... stdby up zfs data/osvprdcollector.opensvc.com/docker@/srv/osvprdcollector.opensvc.com/docker
| |- container#0 .... up docker container osvprdcollector.opensvc.com.container.0@ubuntu:16.04
| |- container#1 .... up docker container osvprdcollector.opensvc.com.container.1@registry.opensvc.com/opensvc/collector_db:build10
| |- container#2 .... up docker container osvprdcollector.opensvc.com.container.2@registry.opensvc.com/opensvc/collector_redis:build1
| |- container#3 .... up docker container osvprdcollector.opensvc.com.container.3@registry.opensvc.com/opensvc/collector_nginx:build1
| '- container#4 .... up docker container osvprdcollector.opensvc.com.container.4@registry.opensvc.com/opensvc/collector_web2py:build10
'- accessory
|- sync#1 .... up zfs of data/osvprdcollector.opensvc.com to nodes
|- sync#1sd .... up zfs 'daily' snapshot data/osvprdcollector.opensvc.com
'- sync#i0 .... up rsync svc config to drpnodes, nodes
Starting a service:
$ sudo mysvc1.opensvc.com start --local
deb1.mysvc1.ip#1 checking 128.0.1.124 availability
deb1.mysvc1.ip#1 ifconfig lo:3 128.0.1.124 netmask 255.255.255.255 up
deb1.mysvc1.ip#1 arping -U -c 1 -I lo -s 128.0.1.124 128.0.1.124
deb1.mysvc1.disk#1 loop /opt/disk1.dd is already up
deb1.mysvc1.disk#2 loop /opt/disk2.dd is already up
deb1.mysvc1.disk#3 vg vgtest is already up
deb1.mysvc1.fs#1 e2fsck -p /dev/vgtest/lvtest1
deb1.mysvc1.fs#1 output:
deb1.mysvc1.fs#1 /dev/vgtest/lvtest1: clean, 18/3072 files, 1534/12288 blocks
deb1.mysvc1.fs#1
deb1.mysvc1.fs#1 mount -t ext4 -o rw /dev/vgtest/lvtest1 /opt/avn/lvtest1
deb1.mysvc1.fs#2 e2fsck -p /dev/vgtest/lvtest2
deb1.mysvc1.fs#2 output:
deb1.mysvc1.fs#2 /dev/vgtest/lvtest2: clean, 13/3072 files, 12286/12288 blocks
deb1.mysvc1.fs#2
deb1.mysvc1.fs#2 mount -t ext4 -o rw /dev/vgtest/lvtest2 /opt/avn/lvtest2
deb1.mysvc1.fs#3 e2fsck -p /dev/loop1
deb1.mysvc1.fs#3 output:
deb1.mysvc1.fs#3 testfs: clean, 13/12824 files, 27111/51200 blocks
deb1.mysvc1.fs#3
deb1.mysvc1.fs#3 mount -t ext4 -o rw /dev/loop1 /opt/avn/lvtest3
deb1.mysvc1.share#0 exportfs -o ro,fsid=0 p145.opensvc.com:/opt/avn/lvtest3
deb1.mysvc1.app#0 exec /bin/true start as user root
deb1.mysvc1.app#0 start done in 0:00:00.001864 - ret 0
Stopping a service:
$ sudo mysvc1.opensvc.com stop --local
deb1.mysvc1.app#0 exec /bin/true stop as user root
deb1.mysvc1.app#0 stop done in 0:00:00.004243 ret 0
deb1.mysvc1.share#0 exportfs -u p145.opensvc.com:/opt/avn/lvtest3
deb1.mysvc1.fs#3 umount /opt/avn/lvtest3
deb1.mysvc1.fs#2 umount /opt/avn/lvtest2
deb1.mysvc1.fs#1 umount /opt/avn/lvtest1
deb1.mysvc1.disk#3 skip 'stop' on standby resource (--force to override)
deb1.mysvc1.disk#2 skip 'stop' on standby resource (--force to override)
deb1.mysvc1.disk#1 skip 'stop' on standby resource (--force to override)
deb1.mysvc1.ip#1 ifconfig lo:3 down