Closed Bug 966048 Opened 10 years ago Closed 10 years ago

Add route to Amazon DNS for try slaves

Categories

(Release Engineering :: General, defect, P2)

Product:
Release Engineering
Component:
General
Platform:
x86
All
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED WONTFIX

People

(Reporter: nthomas, Assigned: nthomas)

References

Details

glandium's s3 cacher isn't as quick as it could be because resolving the S3 endpoint is happening in SCL3 via our in house resolvers. The cacher likely gets directed to resources in us-west-2, which is slower for us-east-1 slaves than something local to them.

He's requesting we add a route to 172.16.0.23 over the internet rather than tunnel.
Added 172.16.0.23/32 to rtb-0f4d506d (us-west-2), and to rtb-eeae4a8b (us-east-1). Also associated some subnets to the latter rtb - probably spots in east were going all the way back to SCL3 and out to S3, rather than out through the amazon internet gateway.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
To be specific, it was the 4 subnets covering 10.134.66.x and 10.134.67.x
I'm still getting timeouts accessing that dns server.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
So it turns out VPCs use a different DNS. So you can revert the change for 172.16.0.23. I'm testing other addresses that may not require routing changes.
Reverted the routing changes, left the subnet associations.
Status: REOPENED → RESOLVED
Closed: 10 years ago → 10 years ago
Resolution: --- → WONTFIX
Component: General Automation → General
You need to log in before you can comment on or make changes to this bug.