try to optimize getting all zfcp controllers

[jreidinger]

Problem

Getting zfcp controllers is very slow call.

Solution

Try thread approach without dbus calls for this specific http call.

Testing

original time on our testing s390

real	0m10.851s
user	0m0.005s
sys	0m0.005s

[coveralls]

Pull Request Test Coverage Report for Build 11243278108

Details

• 0 of 30 (0.0%) changed or added relevant lines in 1 file are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage decreased (-0.07%) to 70.523%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
rust/agama-lib/src/storage/client/zfcp.rs	0	30	0.0%

Files with Coverage Reduction	New Missed Lines	%
rust/agama-lib/src/storage/client/zfcp.rs	1	0.0%

Totals
Change from base Build 11209402098:	-0.07%
Covered Lines:	15836
Relevant Lines:	22455

---

💛 - Coveralls

[imobachgs]

Instead of spawning all those processes, I would suggest giving join_all a try.

[imobachgs]

Tokio has JoinSet, which might be interesting too.

[jreidinger]

thanks for suggestion. I plan to have it as experiment first and this tokio spawn is suggested way in that linkedin rust training, so it is the first that I would like to give a try.

[jreidinger]

and testing shows that my change somehow does not work, but as it is just learning/research project I will put it a bit on hold and focus on planned stuff and return to it probably next week.

[jreidinger]

ok, I cannot sleep without debugging it...so now it is fixed and time is improved to 6.3 second..so already it goes down from 10.8 to 6.3 and I think it can be even better if I parallel also getting wwpns for controller. And then I can start playing with join_all and JoinSet

[jreidinger]

with this additional paralellization time is reduced to 5.3 seconds. So basically for two controllers with 3 LUNS in total it is gets to half of time. I expect that for bigger systems it will be even more significant.