From: Philipp Reisner <philipp.reisner@linbit.com>
To: drbd-dev@lists.linbit.com
Cc: "Montrose, Ernest" <Ernest.Montrose@stratus.com>
Subject: Re: [Drbd-dev] DRBD8: drbd nodes deadlock in WFBitMapT
Date: Tue, 3 Apr 2007 11:34:54 +0200 [thread overview]
Message-ID: <200704031134.55269.philipp.reisner@linbit.com> (raw)
In-Reply-To: <BD7042533C2F8943A6A4257A9E31C45439C948@EXNA.corp.stratus.com>
[-- Attachment #1: Type: text/plain, Size: 1354 bytes --]
Am Montag, 2. April 2007 23:44 schrieb Montrose, Ernest:
> Phil,
> OK...I have learned a tad more since that last email. So before I even
> try the proposed patch here is a way I was able to duplicate the
> problem.
> Perhaps that will help a bit. Here it is with two nodes 'a' and 'b' .
> Suppose
> They are in steady states with UUIDS:
> Xa:0:Ha:HH:1:1:0:1:0:0
> Xb:0:Hb:HH:1:1:0:1:0:0
> Role Secondary/Secondary
>
> 1) Disconnect/detach /dev/drbdX on nodea
> 2) Move Current UUID of nodea to history-UUID of nodea and set
> current_UUID of nodea to 00000000000 with drbdmeta..
> 0:0:Xa:HH:1:1:0:1:0:0
> 3) Now attach and connect /dev/drbdX and the problem will occur
>
> I have attached the logs for my "manufactured" version of the problem
>
Ernest,
You are right, that DRBD should get out of this situation. The attached
patch fixes this. (I will commit it when you confirm that it also fixes
the issue for your)
But I am still asking myself how the CRASHED_PRIMARY got lost. Ernest,
do you still have the log of jerry from Mar 23 13:16:54 ?
I would really like to see the last 30 lines before Mar 23 13:16:54.
Thanks!
-Phil
--
: Dipl-Ing Philipp Reisner Tel +43-1-8178292-50 :
: LINBIT Information Technologies GmbH Fax +43-1-8178292-82 :
: Vivenotgasse 48, 1120 Vienna, Austria http://www.linbit.com :
[-- Attachment #2: fix_i2.diff --]
[-- Type: text/x-diff, Size: 675 bytes --]
Index: drbd_receiver.c
===================================================================
--- drbd_receiver.c (revision 2822)
+++ drbd_receiver.c (working copy)
@@ -1890,7 +1890,7 @@
*rule_nr = 5;
peer = mdev->p_uuid[Bitmap] & ~((u64)1);
- if (self == peer) return -1;
+ if (self == peer && self != ((u64)0)) return -1;
*rule_nr = 6;
for ( i=History_start ; i<=History_end ; i++ ) {
@@ -1901,7 +1901,7 @@
*rule_nr = 7;
self = mdev->bc->md.uuid[Bitmap] & ~((u64)1);
peer = mdev->p_uuid[Current] & ~((u64)1);
- if (self == peer) return 1;
+ if (self == peer && self != ((u64)0)) return 1;
*rule_nr = 8;
for ( i=History_start ; i<=History_end ; i++ ) {
next prev parent reply other threads:[~2007-04-03 9:34 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-03-30 22:01 [Drbd-dev] DRBD8: drbd nodes deadlock in WFBitMapT Montrose, Ernest
2007-04-02 19:59 ` Philipp Reisner
2007-04-02 21:44 ` Montrose, Ernest
2007-04-03 9:34 ` Philipp Reisner [this message]
-- strict thread matches above, loose matches on Subject: below --
2007-04-03 12:49 Montrose, Ernest
2007-04-05 20:41 ` Philipp Reisner
2007-04-05 20:45 Montrose, Ernest
2007-04-06 19:00 Montrose, Ernest
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=200704031134.55269.philipp.reisner@linbit.com \
--to=philipp.reisner@linbit.com \
--cc=Ernest.Montrose@stratus.com \
--cc=drbd-dev@lists.linbit.com \
/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