I have a requirement that I like to bounce off the members first before I start installing things.
There are upto 12 machines in a Food factory, each with an industrial PC running Linux Mint. There is one other pc that can be used as a server in the same facility.Each PC (Fitlets) has Wifi.
A dedicated program operates each of these machines and requires milli second precision. I say that because I can’t afford Network traffic during production. When a machine is stopped for longer than 30 sec’s the report is updated with values. During that time we can sync with the server which has a very light non-time critical function.
The system has been running successfully for a few years now.
I have been asked to automate getting reports to head office on the other side of the country.So getting the reports from the server to HO is my main concern with Syncthing, as the preliminary reading did not shed light on that for me.
The reports do not need to be synced to a schedule, but as the operators turn the power off for cleaning when production finishes picking the time for syncing seems difficult.
The clients can sync with the server as soon as production has stopped for 30 sec’s. Similarly I don’t mind if HO is updated the following day when the power is first turned On.
So, I don’t have a specific question but hoping to prompt someone to reply with a roadmap to achieving this.
I think you are trying to perform brain surgery with a butchers knife. If you have specific requirements, you need to write specific software that does what you need.
Trying to bodge random pieces of software with blu-tack and hoping they will not effect some super sensitive process is very naive.
There are a ton of things in syncthing that are out of your control, which will use network and consume CPU at times that are out of your control.
Furthermore, it also seems that you don’t need bidirectional sync anyway, so I am not sure why you are even looking at syncthing. You can just rsync/sftp the files across whenever you know it’s safe.
In the same direction of what Audrius wrote: To not affect the production machine you’d need some scheduling to start/stop Syncthing when the machine is stopped/started. So you could use the same scheduling to kick off an rsync run. The advantage is that it’s less moving parts and you get an immediate status of whether the transfer is successfully done or not.
For syncing between the server and HO using Syncthing would be a bit more sensible, as you could just keep it running all the time, i.e. no need to worry about scheduling. And you could add more devices e.g. if several pcs need access in HO.
That is the reason I ask questions first.
Many thanks for the responses.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.