fd's are not correctly reopened after a brick reconnection

[xhernandez]

Description of problem:

When a client has many open fd's with active or blocked posix locks and a brick disconnects and reconnects, those fd's are not reopened. This causes partial failures for replicated or dispersed volumes that most probably won't be visible from the application side, only on the logs. However, if all bricks get reconnected, even if this doesn't happen simultaneously, all fd's that have remained open will start to fail with EBADFD on the application side.

The exact command to reproduce the issue:

Using the program provided here, the issue can be seen.

[stale[bot]]

Thank you for your contributions. Noticed that this issue is not having any activity in last ~6 months! We are marking this issue as stale because it has not had recent activity. It will be closed in 2 weeks if no one responds with a comment here.

[stale[bot]]

Thank you for your contributions. Noticed that this issue is not having any activity in last ~6 months! We are marking this issue as stale because it has not had recent activity. It will be closed in 2 weeks if no one responds with a comment here.