These are chat archives for spring-cloud/spring-cloud

15th
Oct 2017
Dave Syer
@dsyer
Oct 15 2017 12:55
Not really a "cloud" question?
It's mostly a matter of taste. Some people like to do both.
Al Grant
@bigalnz
Oct 15 2017 12:56
deleted and will put in right channel
Narayanan Raghavan
@naruraghavan
Oct 15 2017 17:04
@dsyer did you address it to me?
Dave Syer
@dsyer
Oct 15 2017 17:13
No, to Al.
Your question has insufficient context to offer an opinion
Narayanan Raghavan
@naruraghavan
Oct 15 2017 17:13
OH
i can change that
I have the following scenario:
PCF 1.10.x
2 orgs - each runs its own service registry, hysterix dashboard, and config server
Each org also has its own gateway using ZuulProxy
There are services hosted by each org
End user requests come to Org1->Gateway1->service1 and service1 depends on Org2 services and invokes them as Org2->gateway2->service2
Every 10 mins of inactivity,
calls to Org1->Gateway1 throws error 500
Narayanan Raghavan
@naruraghavan
Oct 15 2017 17:19
if look into the logs, I see the following:
Socket Exception: Connection reset
as the root cause of the problem
Running spring cloud services 1.4.3 on PCF
The service registry has the following:
Narayanan Raghavan
@naruraghavan
Oct 15 2017 17:25
Is the 10 mins issue due to “lease expiration"
Narayanan Raghavan
@naruraghavan
Oct 15 2017 17:38
Lease expiration is enabled but self preservation is disabled
Narayanan Raghavan
@naruraghavan
Oct 15 2017 18:02
PCF runs on AWS
Dave Syer
@dsyer
Oct 15 2017 18:41
Sounds like more of a low level (HTTP) problem.
Ask your PCF admins
Also, obviously, upgrade spring cloud to get the latest bug fixes
I'm still not 100% clear where the error is (in the gateway or in the caller).
Narayanan Raghavan
@naruraghavan
Oct 15 2017 18:44
we tried changing AWS ALB timeouts etc. but it did not help
Gateway is where the problem is
caller = browser client
If calls are made every 10 mins, the call will fail first time
with a soket exception
Dave Syer
@dsyer
Oct 15 2017 18:46
There's an open issue in github
Sounds the same to me
Narayanan Raghavan
@naruraghavan
Oct 15 2017 18:46
any remedies
i have already upgraded SCS on PCF to 1.4.3 which is the latest
Dave Syer
@dsyer
Oct 15 2017 18:46
But I didn't really follow. I think the OP actually closed it at an one point and said he fixed it in the LB.
Narayanan Raghavan
@naruraghavan
Oct 15 2017 18:46
but can upgrade code
i saw that and looked at AWS LBs
and adjusted timeouts but did not help
Dave Syer
@dsyer
Oct 15 2017 18:47
Don't know then
Narayanan Raghavan
@naruraghavan
Oct 15 2017 18:48
ok