[Ocfs2-devel] [patch]o2net: fix qs_holds may could not be zero
Zhangyang
zhang.yangB at h3c.com
Mon Sep 11 00:24:56 PDT 2017
Hi all
In our test, We fond that , when the network down, qs->qs_holds may could not be reduce to zero, lead to the node can't do fence.
o2net_idle_timer -> o2quo_conn_err -> qs->qs_holds++, after O2NET_QUORUM_DELAY_MS if qs_holds could be subtract to zero, it could do make_decision.
But if there are many nodes, when one node network down which contains o2net connections may not do o2net_idle_timer at the same time, so when a o2net_node have done nn->nn_still_up, but the qs_holds is not be zero, because other o2net_node have not done nn->nn_still_up.so the first o2net_node will do o2net_idle_timer again, and the qs_holds could be add again. And the qs_holds is global variable, so it formed a loop, the node could not do o2quo_make_decision, because of qs_holds never be zero.
I alter two functions o2quo_conn_up and o2quo_conn_err, take o2quo_set_hold or o2quo_clear_hold under control of the bit map qs_conn_bm
Best Regards,
Yang Zhang
@@ -280,20 +280,24 @@ void o2quo_conn_up(u8 node)
struct o2quo_state *qs = &o2quo_state;
spin_lock(&qs->qs_lock);
-
- qs->qs_connected++;
- mlog_bug_on_msg(qs->qs_connected == O2NM_MAX_NODES,
- "node %u\n", node);
- mlog_bug_on_msg(test_bit(node, qs->qs_conn_bm), "node %u\n", node);
- set_bit(node, qs->qs_conn_bm);
-
- mlog(0, "node %u, %d total\n", node, qs->qs_connected);
-
- if (!test_bit(node, qs->qs_hb_bm))
- o2quo_set_hold(qs, node);
- else
- o2quo_clear_hold(qs, node);
-
+ if (!test_bit(node, qs->qs_conn_bm)) {
+ qs->qs_connected++;
+ mlog_bug_on_msg(qs->qs_connected == O2NM_MAX_NODES,
+ "node %u\n", node);
+ mlog_bug_on_msg(test_bit(node, qs->qs_conn_bm), "node %u\n", node);
+ set_bit(node, qs->qs_conn_bm);
+
+ mlog(0, "node %u, %d total\n", node, qs->qs_connected);
+
+ if (!test_bit(node, qs->qs_hb_bm))
+ o2quo_set_hold(qs, node);
+ else
+ o2quo_clear_hold(qs, node);
+ }
spin_unlock(&qs->qs_lock);
}
@@ -314,13 +318,13 @@ void o2quo_conn_err(u8 node)
node, qs->qs_connected);
clear_bit(node, qs->qs_conn_bm);
+
+ if (test_bit(node, qs->qs_hb_bm))
+ o2quo_set_hold(qs, node);
}
mlog(0, "node %u, %d total\n", node, qs->qs_connected);
- if (test_bit(node, qs->qs_hb_bm))
- o2quo_set_hold(qs, node);
-
spin_unlock(&qs->qs_lock);
}
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oss.oracle.com/pipermail/ocfs2-devel/attachments/20170911/9114b403/attachment.html
More information about the Ocfs2-devel
mailing list