From: teigland@sourceware.org <teigland@sourceware.org>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] cluster/group/daemon app.c
Date: 31 Aug 2006 18:20:51 -0000 [thread overview]
Message-ID: <20060831182051.27530.qmail@sourceware.org> (raw)
CVSROOT: /cvs/cluster
Module name: cluster
Changes by: teigland at sourceware.org 2006-08-31 18:20:51
Modified files:
group/daemon : app.c
Log message:
when we set a recovery event back to the FAIL_BEGIN state, make
sure that we process the event once before processing any new
messages. this is probably a better fix for bz 202635 than I
added previously where we accept messages more liberally i.e. in
X_BEGIN states.
Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/group/daemon/app.c.diff?cvsroot=cluster&r1=1.41&r2=1.42
--- cluster/group/daemon/app.c 2006/08/22 14:36:20 1.41
+++ cluster/group/daemon/app.c 2006/08/31 18:20:51 1.42
@@ -1316,6 +1316,7 @@
event_t *ev, *rev;
node_t *node;
struct nodeid *id, *safe;
+ int rv = 0;
ev = a->current_event;
if (!ev)
@@ -1361,7 +1362,7 @@
list_del(&rev->list);
free_event(rev);
- return 0;
+ return 1;
}
/* Before starting the rev we need to apply the node addition/removal
@@ -1383,7 +1384,7 @@
list_del(&rev->list);
a->current_event = rev;
free_event(ev);
-
+ rv = 1;
} else if (event_state_stopping(a)) {
/* We'll come back through here multiple times until all the
@@ -1397,9 +1398,7 @@
mark_node_stopped(a, rev->nodeid);
list_for_each_entry(id, &rev->extended, list)
mark_node_stopped(a, id->nodeid);
-
- process_current_event(g);
-
+ rv = 1;
} else {
log_group(g, "rev for %d delayed for ev %d %s",
rev->nodeid, ev->nodeid, ev_state_str(ev));
@@ -1411,7 +1410,7 @@
/* FIXME: if the current event is a leave and the leaving node has
failed, then replace the current event with the rev */
- return 0;
+ return rv;
}
static int process_app(group_t *g)
@@ -1428,7 +1427,18 @@
goto out;
rv += ret;
- rv += recover_current_event(g);
+ ret = recover_current_event(g);
+ if (ret > 0) {
+ rv += ret;
+
+ /* it's important that we call process_current_event()
+ when recover_current_event() returns 1 */
+
+ ret = process_current_event(g);
+ if (ret < 0)
+ goto out;
+ rv += ret;
+ }
} else {
/* We only take on a new non-recovery event if there are
no recovery sets outstanding. The new event may be
next reply other threads:[~2006-08-31 18:20 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-08-31 18:20 teigland [this message]
-- strict thread matches above, loose matches on Subject: below --
2006-10-16 15:52 [Cluster-devel] cluster/group/daemon app.c teigland
2006-10-12 14:25 teigland
2006-09-15 20:07 teigland
2006-09-14 20:56 teigland
2006-08-22 14:36 teigland
2006-08-14 21:01 teigland
2006-08-02 20:30 teigland
2006-07-19 19:45 teigland
2006-07-17 18:22 teigland
2006-07-17 17:06 rohara
2006-07-14 20:45 teigland
2006-07-16 7:51 ` Fabio Massimo Di Nitto
2006-07-17 17:08 ` Ryan O'Hara
2006-06-22 21:12 teigland
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20060831182051.27530.qmail@sourceware.org \
--to=teigland@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).