From: teigland@sourceware.org <teigland@sourceware.org>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] cluster/group/gfs_controld lock_dlm.h recover.c
Date: 16 Oct 2006 14:44:03 -0000 [thread overview]
Message-ID: <20061016144403.24169.qmail@sourceware.org> (raw)
CVSROOT: /cvs/cluster
Module name: cluster
Changes by: teigland at sourceware.org 2006-10-16 14:44:02
Modified files:
group/gfs_controld: lock_dlm.h recover.c
Log message:
A node that was just added would incorrectly conclude that the node
after it needed to do first mounter recovery.
Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/group/gfs_controld/lock_dlm.h.diff?cvsroot=cluster&r1=1.19&r2=1.20
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/group/gfs_controld/recover.c.diff?cvsroot=cluster&r1=1.19&r2=1.20
--- cluster/group/gfs_controld/lock_dlm.h 2006/10/13 20:00:02 1.19
+++ cluster/group/gfs_controld/lock_dlm.h 2006/10/16 14:44:02 1.20
@@ -149,6 +149,7 @@
int first_mount_pending_stop;
int first_mounter;
int first_mounter_done;
+ int global_first_recover_done;
int emulate_first_mounter;
int wait_first_done;
int low_nodeid;
--- cluster/group/gfs_controld/recover.c 2006/10/13 20:00:02 1.19
+++ cluster/group/gfs_controld/recover.c 2006/10/16 14:44:02 1.20
@@ -822,9 +822,22 @@
goto out;
}
+ /* when we received our journals, no one was flagged with OPT_RECOVER
+ which means no first mounter recovery is needed or is current */
+
+ if (mg->global_first_recover_done) {
+ log_group(mg, "assign_journal: global_firsts_recover_done");
+ goto out;
+ }
+
/* no one has done kernel mount successfully and no one is doing first
mounter recovery, the new node gets to try first mounter recovery */
+ log_group(mg, "kernel_mount_done %d kernel_mount_error %d "
+ "first_mounter %d first_mounter_done %d",
+ mg->kernel_mount_done, mg->kernel_mount_error,
+ mg->first_mounter, mg->first_mounter_done);
+
log_group(mg, "assign_journal: memb %d gets OPT_RECOVER", new->nodeid);
new->opts |= MEMB_OPT_RECOVER;
@@ -1007,6 +1020,7 @@
struct mg_member *memb, *memb2;
struct gdlm_header *hd;
int *ids, count, i, nodeid, jid, opts;
+ int current_first_recover = 0;
hd = (struct gdlm_header *)buf;
@@ -1048,8 +1062,16 @@
else if (opts & MEMB_OPT_SPECT)
memb->spectator = 1;
}
+
+ if (opts & MEMB_OPT_RECOVER)
+ current_first_recover = 1;
}
+ /* FIXME: use global_first_recover_done more widely instead of
+ as a single special case */
+ if (!current_first_recover)
+ mg->global_first_recover_done = 1;
+
process_saved_mount_status(mg);
/* we delay processing any options messages from new mounters
next reply other threads:[~2006-10-16 14:44 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-10-16 14:44 teigland [this message]
-- strict thread matches above, loose matches on Subject: below --
2006-10-23 15:44 [Cluster-devel] cluster/group/gfs_controld lock_dlm.h recover.c teigland
2006-12-19 1:42 rpeterso
2006-12-19 1:46 rpeterso
2006-12-19 1:48 rpeterso
2006-12-19 17:06 teigland
2006-12-19 17:07 teigland
2006-12-19 17:07 teigland
2007-09-04 19:22 teigland
2007-09-04 19:27 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=20061016144403.24169.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).