From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Philipp Reisner To: "Montrose, Ernest" Subject: Re: [Drbd-dev] Stuck in WFBitMapS again Date: Mon, 31 Mar 2008 16:43:51 +0200 References: <342BAC0A5467384983B586A6B0B3767107E8A3F8@EXNA.corp.stratus.com> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200803311643.51501.philipp.reisner@linbit.com> Cc: drbd-dev@lists.linbit.com List-Id: Coordination of development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Am Mittwoch, 13. Februar 2008 00:55:51 schrieb Montrose, Ernest: > It looks like we still have a race condition while finishing syncing and > pausing/resuming syncing at the same time. > > What's happening is that we finish syncing. > The secondary side goes to Connected at the end of > drbd_resync_finished() > > The Primary side is still in drbd_resync_finished(). While still there, > > the syncer isp flag gets set and unset due to a higher priority device > syncer getting paused/resumed. > > Now the Primary side sends "Inconsistent" state to the Secondary since > we are still in drbd_resync_finished() > and have not gone Connected and UptoDate yet. But the other side is > connected and UpToDate though. > > The secondary goes WFBitMapS via drbd_sync_handshake() > > In the mean time the Primary finally goes Connected and UptoDate. > > The Secondary is stuck in WFBitMapS. > > I include a patch that may be the fix for this. Any other ideas? > > Thanks. > EM-- Hi Ernest, I know it is quite some time now, but I had to remove this patch. It breaks manual split brain resolution. Commit ab7facf4e17d98558d2569ccf954f2c64eb7b650 That was the wrong way of fixing this. Do you have something with logfiles of that event -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 :