condor_advertise can write to the COLLECTOR, but the data in the collector is stored only in memory, so this would be a runtime-only key-value store. You would have to re-advertise everything each time the collector was restarted. condor_submit and condor_qedit can write to Job ads in the Schedd, and these are persisted to disk and reloaded. But you canât alter job attributes of running jobs - or rather you can, but the running job will never see the changes and may overwrite them depending on the attribute. -tj From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx>
On Behalf Of don_vanchos Hello, This video https://youtu.be/A_YKDRD_LzE?t=908 says that HTCondor can be used as NoSql database. The announcer explains: "and what you can do is just put these attributes right into the job. ... and you can query on these ..." Nevertheless, I do not understand. What methods should I use to use HTCondor as key-value storage? Can I access this storage and change values while the task has RUNNING state? On the other hand can I use `condor_advertise` as API to reliable data storage? If so, where does the Collector really store this data? In the daemon configuration files? -- Sincerely yours, |