Hi Sergii,
Thanks for the quick response , I see your point .
The issue from my side is that our logic from the runner implementation side, causes the runner to reach Stopped state in 2 different flows .(the reason for the
stopped state is the logic implemented in CHE runner.)
For the sake of the example, let's call them "Flow A" and " Flow B".
Now if we look at this case from the Clients side , he issued a run request and the process is stopped.
How can the client differentiate between runner " Flow A " or " Flow B" ?
My initial idea was to use the Failed state, since IMHO I need an additional state for the process .
But then I saw that any action that can be executed on a given process is allowed only on NEW, STOPED, CANCELED, RUNNING states (the notorious "if" statement
we discuss in the previous email).
And following your response bellow, I cannot and will not be able to use the failed state following your vision of things . that's an issue from my side, since
I need an additional state of the process.
Hence I need your help with this situation – how can we address this ?
I can think of a number of solutions but I rather you propose the first option , since I wish to be aligned with your vision.
BR
--Dany
From: Sergii Kabashniuk [mailto:skabashnyuk@xxxxxxxxxxx]
Sent: Monday 08 February 2016 09:41
To: Shapiro, Dany <dany.shapiro@xxxxxxx>
Cc: che developer discussions <che-dev@xxxxxxxxxxx>; Milstein, Tahel <tahel.milstein@xxxxxxx>; Sharafy, Tareq <tareq.sharafy@xxxxxxx>; Fishman, Roy <roy.fishman@xxxxxxx>; Pronyakova, Zoya <zoya.pronyakova@xxxxxxx>; Fonio, Itai <itai.fonio@xxxxxxx>; Vagmayster,
Nachshon <nachshon.vagmayster@xxxxxxx>; Vitalii Parfonov <vparfonov@xxxxxxxxxxx>
Subject: Re: Faild status in CHE core.
Hello Dany
Our vision for this situation is the following:
Runner implementation should handle all errors on start and clean all used resources.
So "Failed" state is just a state for information use only.
Should be no need to stop "failed" runner because it doesn't use any resources.
On Sun, Feb 7, 2016 at 4:04 PM, Shapiro, Dany <dany.shapiro@xxxxxxx> wrote:
Hi Sergii,
Hope you remember the meeting we had (you zoya and me ) regarding a few issues we found in che .
One of the issues we discussed (that I'm wondering about its status )was as following:
When a specific runner implementation throw an Runner exception , this puts the status of the newly created process in Failed.
The issue is as following –from the core side when someone wishes to stop/delete that process (with failed as its status ), it cannot be done .
The reason for this is that there is an "if” statement that doesn’t allow stop operation of processes with the status Failed.
It's Equivalent to a situation that doesn’t allow recovery from error state, so the process will be there for ever .
Like I mentioned before, I was wondering was there any progress regarding that issue. During the last meeting you told me you will talk with the developers and get back to me .
we haven’t spoken since, hence I would like to know if you have any updates ?
Thanks
Dany