From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 24075 invoked by alias); 9 Dec 2009 19:29:05 -0000 Received: (qmail 24060 invoked by uid 9664); 9 Dec 2009 19:29:04 -0000 Date: Wed, 09 Dec 2009 19:29:00 -0000 Message-ID: <20091209192904.24058.qmail@sourceware.org> From: mbroz@sourceware.org To: lvm-devel@redhat.com, lvm2-cvs@sourceware.org Subject: LVM2 ./WHATS_NEW lib/metadata/metadata.c Mailing-List: contact lvm2-cvs-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: lvm2-cvs-owner@sourceware.org X-SW-Source: 2009-12/txt/msg00024.txt.bz2 CVSROOT: /cvs/lvm2 Module name: LVM2 Changes by: mbroz@sourceware.org 2009-12-09 19:29:04 Modified files: . : WHATS_NEW lib/metadata : metadata.c Log message: Allow manipulation with precommited metadata even when a PV is missing. The new recovery code first tries to repair LV and then removes failed PV from VG. It means that during operation there can be VG with PV missing, and vg_read code handles it like not consistent VG. We already allows returning "inconsistent" commited metadata, for mirror repair we need this for precommited too. (The suspend call prepares precommited metadata to inactive table on other cluster nodes.) "Inconsistent" here means - correct metadata, just with some metadata areas not found (obviously on missing or failed PVs). Patches: http://sourceware.org/cgi-bin/cvsweb.cgi/LVM2/WHATS_NEW.diff?cvsroot=lvm2&r1=1.1348&r2=1.1349 http://sourceware.org/cgi-bin/cvsweb.cgi/LVM2/lib/metadata/metadata.c.diff?cvsroot=lvm2&r1=1.300&r2=1.301 --- LVM2/WHATS_NEW 2009/12/09 19:01:27 1.1348 +++ LVM2/WHATS_NEW 2009/12/09 19:29:04 1.1349 @@ -1,5 +1,6 @@ Version 2.02.57 - ==================================== + Allow use precommited metadata when a PV is missing. Add memlock information to do_lock_lv debug output. Never use distributed lock for LV in non-clustered VG. Allow implicit lock conversion in pre/post callbacks on local node. --- LVM2/lib/metadata/metadata.c 2009/12/03 19:18:34 1.300 +++ LVM2/lib/metadata/metadata.c 2009/12/09 19:29:04 1.301 @@ -2481,6 +2481,7 @@ int inconsistent = 0; int inconsistent_vgid = 0; int inconsistent_pvs = 0; + int inconsistent_seqno = 0; unsigned use_precommitted = precommitted; unsigned saved_handles_missing_pvs = cmd->handles_missing_pvs; struct dm_list *pvids; @@ -2556,6 +2557,7 @@ /* FIXME Also ensure contents same - checksum compare? */ if (correct_vg->seqno != vg->seqno) { inconsistent = 1; + inconsistent_seqno = 1; if (vg->seqno > correct_vg->seqno) { vg_release(correct_vg); correct_vg = vg; @@ -2679,6 +2681,7 @@ /* FIXME Also ensure contents same - checksums same? */ if (correct_vg->seqno != vg->seqno) { inconsistent = 1; + inconsistent_seqno = 1; if (!_update_pv_list(cmd->mem, &all_pvs, vg)) { vg_release(vg); vg_release(correct_vg); @@ -2706,6 +2709,16 @@ if (use_precommitted) { log_error("Inconsistent pre-commit metadata copies " "for volume group %s", vgname); + /* FIXME: during repair, there is inconsistent flag set because some metadata areas + * are missing (on missing PVs). Code should create list of missing PVs, compare it + * with PV marked missing in metadata and if equals, use it as consistent vg. + * For now, return precommited metadata if remainng seq match here to allow + * preloading table in suspend call. + */ + if (!inconsistent_seqno) { + *consistent = 0; + return correct_vg; + } vg_release(correct_vg); return NULL; }