-
Notifications
You must be signed in to change notification settings - Fork 49
Conversation
Codecov Report
@@ Coverage Diff @@
## master #132 +/- ##
==========================================
- Coverage 35.81% 33.34% -2.48%
==========================================
Files 65 61 -4
Lines 7399 7066 -333
==========================================
- Hits 2650 2356 -294
+ Misses 4445 4422 -23
+ Partials 304 288 -16
|
d487585
to
acde9fb
Compare
Pull Request Test Coverage Report for Build 1187
💛 - Coveralls |
@@ -661,14 +671,29 @@ func prepareNetworkDriver(cfg Config, c *runtimeTypes.Container) error { | |||
"--batch-size", strconv.Itoa(cfg.batchSize), | |||
} | |||
|
|||
// This blocks, and ignores kills. | |||
if !c.TitusInfo.GetIgnoreLaunchGuard() { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If IgnoreLaunchGuard is true, why would we want to skip this step?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If IgnoreLaunchGuard is true, we should be using the kill initiated state, so we should never have two tasks using the same interface for different security groups.
if exitErr, ok := e.(*exec.ExitError); ok { | ||
c.SetupCommandStatus <- exitErr | ||
} else { | ||
log.Error("Could not handle exit error of setup command: ", e) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should we push e through the channel?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That was an error.
fslocker/fslocker.go
Outdated
@@ -118,8 +120,9 @@ func (locker *FSLocker) ExclusiveLock(path string, timeout *time.Duration) (*Exc | |||
} | |||
|
|||
// SharedLock tries to get an exclusive Lock on the path |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Comment should be s/exclusive/shared/
@@ -51,10 +51,15 @@ var AllocateNetwork = cli.Command{ // nolint: golint | |||
Usage: "How long to wait for security groups to converge, in seconds", | |||
Value: 10 * time.Second, | |||
}, | |||
cli.DurationFlag{ | |||
Name: "wait-for-sg-lock-timeout", | |||
Usage: "How long to wait for other users, if the SG is in use", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In what scenario would we now expect contention on the same SG? If the SG should be released prior to signaling master of completion, is this just waiting in unexpected/error scenarios?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This timeout will never get hit in the case of the same set of SGs. Only when changing SGs. It's basically, because v2 engine is going to stick around for the forseeable, and we need a solution for that. This is kind of like launchguard-lite.
1b1aa5e
to
907574a
Compare
907574a
to
edcb4fd
Compare
No description provided.