ceph-fuse: start up log on parent process before shutdown #12347
Conversation
Otherwise, we hit an assert in the Ceph context and logging teardown. Fixes: http://tracker.ceph.com/issues/18157 Signed-off-by: Greg Farnum <gfarnum@redhat.com>
|
This is the obvious quick fix — we stop the logging, fork, and then the parent shuts down without restarting the log, then asserts. I'm missing some context here, though. Why do we stop the log @liewegas? |
|
@tchaikov, this is from one of your PRs, any comments? |
|
I tested this by, you know, running ceph-fuse and observing no more crashes. ;) |
|
It's a bit silly to open the log again before shutdown, but I think it's safe. We don't log anything from the parent thread. |
| @@ -295,6 +295,8 @@ int main(int argc, const char **argv, const char *envp[]) { | |||
| //cout << "child done" << std::endl; | |||
| return r; | |||
| } else { | |||
| if (restart_log) | |||
There was a problem hiding this comment.
maybe we'd better use global_init_prefork() before the fork() and call global_init_postfork_start() and global_init_postfork_finish() after fork(). also, why don't we use the Preforker here?
There was a problem hiding this comment.
If you want to try and put together something better that'd be welcome. But I think we need a fix pretty much now and I'm not familiar enough with the init code to make those kinds of changes, @tchaikov.
There was a problem hiding this comment.
@gregsfortytwo sure, i am on it.
|
ceph-fuse's fork hackery predates Preforker
|
Otherwise, we hit an assert in the Ceph context and logging teardown.
Fixes: http://tracker.ceph.com/issues/18157
Signed-off-by: Greg Farnum gfarnum@redhat.com