Whamcloud - gitweb
LU-12292 lnet: keep health even if recovery failed
Don't decrement the interface's health value when recovery
message fails. If we've already determined that an interface
is unhealthy, there is no need to continue decrementing
it's health every 1 second. It'll take too long to come back
into service when it becomes healthy.
Clean up where health is decremented in order not to have
repetitive decrements. No need to decrement in lnet_notify()
because in order for the LND to call this an existing transmit
must've failed. This means a message has already failed which
will result in the health being decremented.
When a recovery send fails make sure to flag the recovery as
failed because there is no reply expected in this case.
Test-parameters: trivial
Signed-off-by: Amir Shehata <ashehata@whamcloud.com>
Change-Id: Ifb3500a77a5a5be51e7079269c8ddba85ed0c2a7
Reviewed-on: https://review.whamcloud.com/36921
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>