Description of problem: Need API for identifying the Randomly selected HSM selected by the engine to perform the SDM_Copy_Data operation with new Task infrastructure without Master File system Persistency Connected to the new Task Infrastructure without Master File System Persistence
This is a blocker on HSM automation, which could be a major risk to the version. We need this to be possible urgently. Please provide workaround or fix.
(In reply to Yaniv Dary from comment #1) > This is a blocker on HSM automation, which could be a major risk to the > version. > We need this to be possible urgently. Please provide workaround or fix. It is by no means not a blocker. They can use the JSON-RPC client of VDSM and queries all hosts in parallel to quickly and swiftly get an answer for where the task is.
The engine log also specify the host where the command run : CopyVolumeDataVDSCommand(HostName = 10.35.1.102, CopyVolumeDataVDSCommandParameters:{runAsync='true', hostId='99d26a0b-acaa-40e7-829b-c7153d55c83f', storageDomainId='null', jobId='e52a03af-bab1-4dfa-97c1-5d4e7829cb5c', srcInfo='VdsmImageLocationInfo [storageDomainId=0b86fe39-423d-4682-b263-bb0f06dd0594, imageGroupId=efc453ff-8c49-42d4-92ba-ed971cf7ccbd, imageId=e2d51152-abea-40ed-a2d0-05199429b6a5, generation=null]', dstInfo='VdsmImageLocationInfo [storageDomainId=51871a7a-75cf-4e50-a223-8f95f4107274, imageGroupId=efc453ff-8c49-42d4-92ba-ed971cf7ccbd, imageId=e2d51152-abea-40ed-a2d0-05199429b6a5, generation=0]', collapse='false'}), log id: 4781252c We can add additional logs as a first phase if it will help.
Raz, please review and lower severity, if you think this is a valid workaround.
As a start, let's add additional logs as a first phase like Fred suggested. It will make it much easier for us (QE) to test
Can you please open a bug on the enhancement you decided on to the logs?
We decided to get the relevant information via REST API /api/jobs/{job_id}/steps and in the steps collection, each step will include the executor hsm as part of the step object. Fred, Is that correct?
Yes. It will need infra approval. Liron will talk with them.
Usage of the api - 1. Have a 4.1 DC in status up with one host and two storage domains 2. Create an image disk 3. Move the disk issuing GET request to /ovirt-engine/api/jobs/JOB_ID/steps/ and to /ovirt-engine/api/jobs/JOB_IDsteps/STEP_ID will allow to see the host performing the step for SPDM steps (if exists) under<step><execution_host> <step href= ""> <actions> <link href= "/ovirt-engine/api/jobs/f3bce79c-3a39-468f-8ec9-b31758af9ff5/steps/066907c6-99d4-4803-b295-b9dd82602062/end" rel="end"/> </actions> <description>Copying Volume</description> rel="statistics"/> <status>finished</status> <type>unknown</type> <execution_host href= "/ovirt-engine/api/hosts/c717ea8e-e839-471a-ab24-df19aa4bca33" id="c717ea8e-e839-471a-ab24-df19aa4bca33"/> <job href= /> <parent_step "/> </step>
Verified on ovirt-engine-4.1.0.2-0.2.el7.noarch The execution host in the API compared to the HostName in engine.log is the same.