Right now, we have no existing bots that are running on "Plus" devices, but we want to run iOS 11 bots with "Plus" devices.

Mike, I don't know how to disable tests only on "Plus" devices, do you prefer disabling this test everywhere or just on iOS 11?

I guess 11 is better, since it still gives coverage (even though the filter isn't accurate of the problem). I'm fine with whatever you decide.

The CL to disable the test on iOS 11.
https://chromium-review.googlesource.com/c/598490

Now the FYI EarlGrey-simulator is running EG suites on iPhone 6 Plus, iOS 10: https://uberchromegw.corp.google.com/i/internal.bling.fyi/builders/EarlGrey-simulator.

Here are failure examples:

https://uberchromegw.corp.google.com/i/internal.bling.fyi/builders/EarlGrey-simulator/builds/8965
https://uberchromegw.corp.google.com/i/internal.bling.fyi/builders/EarlGrey-simulator/builds/8964

NOTE: this test is not failing 3 times in a row, instead, it always fails on the first run, and succeeds when re-try.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7f5284e388f4ef8349e9b5c0fe3b70a1d21b347d

commit 7f5284e388f4ef8349e9b5c0fe3b70a1d21b347d
Author: Yuke Liao <liaoyuke@chromium.org>
Date: Wed Aug 02 19:03:30 2017

Disable testTopSitesTileAfterRedirect on iOS 11.

testTopSitesTileAfterRedirect fails on "Plus" devices, and this CL
disables it on iOS to unblock running the test case on iPhone 6 Plus,
iOS 11.

Bug:  751224 
Change-Id: Iebfac0c16c9f01a68169a19f439e04e828ba8de0
Reviewed-on: https://chromium-review.googlesource.com/598490
Reviewed-by: Chris Pickel <sfiera@chromium.org>
Commit-Queue: Yuke Liao <liaoyuke@chromium.org>
Cr-Commit-Position: refs/heads/master@{#491458}
[modify] https://crrev.com/7f5284e388f4ef8349e9b5c0fe3b70a1d21b347d/ios/chrome/browser/ntp_tiles/ntp_tiles_egtest.mm

I don't think this is Plus-device specific. I see some current failures on iPad Air 2:
https://build.chromium.org/p/chromium.fyi/builders/EarlGreyiOS/builds/30326/steps/ios_chrome_integration_egtests%20%28iPad%20Air%202%20iOS%2010.0%29

Anyway, I think I have a fix.

The problem is running this test after the other test in the suite. If it runs by itself, it passes (which is why the build is green).

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d199756965ab0a7181c89beca16ba5f745b4f69d

commit d199756965ab0a7181c89beca16ba5f745b4f69d
Author: Mike Baxley <baxley@chromium.org>
Date: Fri Aug 04 21:39:13 2017

Call superclass tearDown in NTPTilesTest.

This test overrides the per test tearDown, but had failed to call the
superclass.

Bug:  751224 , 747369 
Change-Id: I895c9115439c06e8630013d4a958c3535b1a3047
Reviewed-on: https://chromium-review.googlesource.com/602682
Reviewed-by: Eugene But <eugenebut@chromium.org>
Commit-Queue: Mike Baxley <baxley@chromium.org>
Cr-Commit-Position: refs/heads/master@{#492128}
[modify] https://crrev.com/d199756965ab0a7181c89beca16ba5f745b4f69d/ios/chrome/browser/ntp_tiles/ntp_tiles_egtest.mm