Services can be started multiple times - does that make sense? (closed)

Added by ragnar-gd over 4 years ago

Starting shard_start on my windows-server twice in a row, almost all services can be started again. And again.
Some throw errors, which can be "ignore"-clicked, but they do start, and remain.
I don't know if this is even possible on linux, as i don't have it running on linux yet.

I did not even try to connect with a client, less identify the error-messages.

Of course, this could be simply stupid, so saying "You don't jump off a cliff just because you can, wouldn't you? So don't do that ..." should suffice.

But, just in case: Is there a hidden reason you can, or where it would make sense?

From my professional life, i know how to transform simple executables into services via commandlinetool "sc.exe", from the windows server ressourcekit, so you can run services without being logged on. That also allows for some tricks, like remote-starting services via "net start <servicename>", interdependencies, startup-chains, automatic restart etc.

Did anyone ever consider this? Are there good reasons against it?
(It would require separating the log-windows from the services themselves, of course, but that would neither necessarily be a bad thing in itself, nor too difficult)


Replies (2)

RE: Services can be started multiple times - does that make sense? - Added by sfb over 4 years ago

ragnar,

It does make sense since you could in theory run multiple shards on a given box. There are only a couple services which should only be run once or once per box. One of the reasons things are not started as a Windows service is that it is uncommon to run the services on a Windows box. Historically only content developers and programmers who needed a temporary environment for testing would run on Windows.

HTH

Thanks,
sfb
/s

RE: Services can be started multiple times - does that make sense? - Added by ragnar-gd over 4 years ago

sfb,

your answer is reasonable, as well regarding windows as a testing-environment, as regarding running multiple shards. Thank you!

I'm already preparing the linux-environment, to be able to present results of our work here to the public as soon as possible (aiming at about three to six months from now on).

Well, the problem can be handled by thinking before starting batches, so i'll use my spare time for more pressing issues... :)

Nonetheless, i'll visit http://dev.ryzom.com/wiki/ryzom/ServiceArchitecture one more time, and see, if i can make a meaningful picture.

regards,
Ragnar G.D.

(1-2/2)