Pass in network IDs during restart#38580
Conversation
|
@fcrisciani PTAL |
|
Could you provide a description of the change (also in the commit message)? |
fcrisciani
left a comment
There was a problem hiding this comment.
Add commit message with some details on the change, but LGTM
selansen
left a comment
There was a problem hiding this comment.
LGTM.
Like @thaJeztah mentioned , pls add description.
|
Just looked into failed test cases. They don't seem related to this. But both Windows CIs are failing. Pls confirm the same. |
c27c250 to
5f2c07b
Compare
Codecov Report
@@ Coverage Diff @@
## master #38580 +/- ##
=========================================
+ Coverage 36.6% 37.01% +0.4%
=========================================
Files 608 610 +2
Lines 45224 45751 +527
=========================================
+ Hits 16555 16933 +378
- Misses 26383 26490 +107
- Partials 2286 2328 +42 |
5f2c07b to
34fa7ef
Compare
|
investigating why test "TestStartReturnCorrectExitCode" fails on RS1... |
|
|
|
so after some investigation of failed tests on RS1 I find out that problem is that in new version hcsshim GetContainers changed returned error this happened after this commit: moby/daemon/graphdriver/windows/windows.go Line 276 in 5ec3138 more specifically condition |
|
@andrey-ko will it be possible to capture the findings around GetContainers above in a fresh issue? |
|
ping @jstarks @jhowardmsft PTAL ^^ wondering if we can get utility functions in hcsshim to prevent such breaks (thinking of a Lines 3 to 11 in d48392a Lines 32 to 42 in d48392a |
|
(removed comment I didn't read the diff correctly) But, perhaps a causer interface would still make sense |
for windows all networks are re-populated in the store during network controller initialization. In current version it also regenerate network Ids which may be referenced by other components and it may cause broken references to a networks. This commit avoids regeneration of network ids. Signed-off-by: Andrey Kolomentsev <[email protected]>
34fa7ef to
e017717
Compare
|
What's the status here? Seems like it's all good, tests are green. |
Other issue discussed above is actually unrelated to this PR, so we can merge it |
|
ok, let's get this merged then 👍 thanks! |
|
@ddebroy @andrey-ko there's some existing issues about restart-policies and windows containers; this one looks like it could be resolved by this fix; could you check? #29832 also (but they seem different, so likely unrelated); #33542, #35430 |
for windows all networks are re-populated in the store during network controller initialization. In current version it also regenerate network Ids which may be referenced by other components and it may cause broken references to networks. This commit avoids regeneration of network ids.