lvm2-commits October 2017

lvm2-commits@lists.fedorahosted.org

8 participants
83 discussions

master - raid: fix validation check for raid0 aegment data_offset member
by Heinz Mauelshagen 09 Oct '17

09 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=85604c0bf10e14547427c… Commit: 85604c0bf10e14547427c428d7b8cb3f57cdbeac Parent: 5f13e33d541f7af77f586ac55edfed336ad8dcc1 Author: Heinz Mauelshagen <heinzm(a)redhat.com> AuthorDate: Mon Oct 9 15:21:51 2017 +0200 Committer: Heinz Mauelshagen <heinzm(a)redhat.com> CommitterDate: Mon Oct 9 15:21:51 2017 +0200 raid: fix validation check for raid0 aegment data_offset member Commit 2f754b73ffdaa95af4a6d703e31b76d4efbf7a36 missed one. --- lib/metadata/merge.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/lib/metadata/merge.c b/lib/metadata/merge.c index 5017cf1..711e7e7 100644 --- a/lib/metadata/merge.c +++ b/lib/metadata/merge.c @@ -119,7 +119,7 @@ static void _check_raid0_seg(struct lv_segment *seg, int *error_count) raid_seg_error_val("non-zero min recovery rate", seg->min_recovery_rate); if (seg->max_recovery_rate) raid_seg_error_val("non-zero max recovery rate", seg->max_recovery_rate); - if ((seg->lv->status & LV_RESHAPE_DATA_OFFSET) || seg->data_offset) + if ((seg->lv->status & LV_RESHAPE_DATA_OFFSET) || seg->data_offset > 1) raid_seg_error_val("data_offset", seg->data_offset); }

1 0

master - lvcreate: fix region size on striped RaidLVs
by Heinz Mauelshagen 09 Oct '17

09 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=5f13e33d541f7af77f586… Commit: 5f13e33d541f7af77f586ac55edfed336ad8dcc1 Parent: 2f754b73ffdaa95af4a6d703e31b76d4efbf7a36 Author: Heinz Mauelshagen <heinzm(a)redhat.com> AuthorDate: Mon Oct 9 14:34:10 2017 +0200 Committer: Heinz Mauelshagen <heinzm(a)redhat.com> CommitterDate: Mon Oct 9 14:35:06 2017 +0200 lvcreate: fix region size on striped RaidLVs Creating striped RaidLVs with lv size not divisible by region size caused the region size to be adjusted: # lvcreate --type raid5 -n region_check.32.00m_3 -i 3 -L 1g --nosync -R 32.00m raid_sanity Using default stripesize 64.00 KiB. Rounding size 1.00 GiB (256 extents) up to stripe boundary size <1.01 GiB(258 extents). WARNING: New raid5 won't be synchronised. Don't read what you didn't write! Using reduced mirror region size of 8.00 MiB Logical volume region_check.32.00m_3 created. Fix by not imposing "mirror" constraints on "raid". Resolves: rhbz1404007 --- lib/metadata/lv_manip.c | 17 ++++++++++++----- 1 files changed, 12 insertions(+), 5 deletions(-) diff --git a/lib/metadata/lv_manip.c b/lib/metadata/lv_manip.c index 1851783..719fcba 100644 --- a/lib/metadata/lv_manip.c +++ b/lib/metadata/lv_manip.c @@ -7536,11 +7536,18 @@ static struct logical_volume *_lv_create_an_lv(struct volume_group *vg, status |= LV_NOTSYNCED; } - lp->region_size = adjusted_mirror_region_size(vg->cmd, - vg->extent_size, - lp->extents, - lp->region_size, 0, - vg_is_clustered(vg)); + if (seg_is_raid(lp)) { + /* Value raid target constraint */ + if (lp->region_size > (uint64_t)vg->extent_size * lp->extents) { + log_error("Cannot create RAID LV with region size larger than LV size."); + return NULL; + } + } else + lp->region_size = adjusted_mirror_region_size(vg->cmd, + vg->extent_size, + lp->extents, + lp->region_size, 0, + vg_is_clustered(vg)); } else if (pool_lv && seg_is_thin_volume(lp)) { if (!lv_is_thin_pool(pool_lv)) { log_error("Logical volume %s is not a thin pool.",

1 0

master - raid: fix validation checks for segment data_offset member
by Heinz Mauelshagen 09 Oct '17

09 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=2f754b73ffdaa95af4a6d… Commit: 2f754b73ffdaa95af4a6d703e31b76d4efbf7a36 Parent: 554a761db2acf2216eadf0279a390ac924038286 Author: Heinz Mauelshagen <heinzm(a)redhat.com> AuthorDate: Mon Oct 9 14:00:46 2017 +0200 Committer: Heinz Mauelshagen <heinzm(a)redhat.com> CommitterDate: Mon Oct 9 14:01:23 2017 +0200 raid: fix validation checks for segment data_offset member Commit 222e1e3acee399b0acf31565f784716d67c465a8 was not valuing special case of data_ofset member equal to 1. --- lib/metadata/merge.c | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/metadata/merge.c b/lib/metadata/merge.c index b0adc5a..5017cf1 100644 --- a/lib/metadata/merge.c +++ b/lib/metadata/merge.c @@ -143,7 +143,7 @@ static void _check_raid1_seg(struct lv_segment *seg, int *error_count) raid_seg_error("no meta areas"); if (seg->stripe_size) raid_seg_error_val("non-zero stripe size", seg->stripe_size); - if ((seg->lv->status & LV_RESHAPE_DATA_OFFSET) || seg->data_offset) + if ((seg->lv->status & LV_RESHAPE_DATA_OFFSET) || seg->data_offset > 1) raid_seg_error_val("data_offset", seg->data_offset); _check_raid_region_recovery(seg, error_count); } @@ -170,7 +170,7 @@ static void _check_raid45610_seg(struct lv_segment *seg, int *error_count) /* END: checks applying to any raid4/5/6/10 */ if (seg->lv->status & LV_RESHAPE_DATA_OFFSET) { - if (seg->data_offset & (seg->lv->vg->extent_size - 1)) + if (seg->data_offset > 1 && (seg->data_offset & (seg->lv->vg->extent_size - 1))) raid_seg_error_val("data_offset", seg->data_offset); } else if (seg->data_offset) raid_seg_error_val("data_offset", seg->data_offset);

1 0

master - raid: return previous reshape space allocation properly
by Heinz Mauelshagen 09 Oct '17

09 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=554a761db2acf2216eadf… Commit: 554a761db2acf2216eadf0279a390ac924038286 Parent: f005a6e7922d9a95eadfbd433714c80301eba794 Author: Heinz Mauelshagen <heinzm(a)redhat.com> AuthorDate: Mon Oct 9 13:53:27 2017 +0200 Committer: Heinz Mauelshagen <heinzm(a)redhat.com> CommitterDate: Mon Oct 9 13:55:01 2017 +0200 raid: return previous reshape space allocation properly Fix returning previous allocation of reshape space. --- lib/metadata/raid_manip.c | 8 +++++--- 1 files changed, 5 insertions(+), 3 deletions(-) diff --git a/lib/metadata/raid_manip.c b/lib/metadata/raid_manip.c index 885a1ca..691ff58 100644 --- a/lib/metadata/raid_manip.c +++ b/lib/metadata/raid_manip.c @@ -1692,10 +1692,10 @@ static int _lv_alloc_reshape_space(struct logical_volume *lv, static int _lv_free_reshape_space_with_status(struct logical_volume *lv, enum alloc_where *where_it_was) { uint32_t total_reshape_len; + enum alloc_where where; struct lv_segment *seg = first_seg(lv); if ((total_reshape_len = _reshape_len_per_lv(lv))) { - enum alloc_where where; /* * raid10: * @@ -1743,9 +1743,11 @@ static int _lv_free_reshape_space_with_status(struct logical_volume *lv, enum al return_0; lv->status &= ~LV_RESHAPE_DATA_OFFSET; + } else + where = alloc_none; - } else if (where_it_was) - *where_it_was = alloc_none; + if (where_it_was) + *where_it_was = where; lv->status &= ~LV_RESHAPE;

1 0

master - man: regenerate
by Alasdair Kergon 06 Oct '17

06 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=f005a6e7922d9a95eadfb… Commit: f005a6e7922d9a95eadfbd433714c80301eba794 Parent: a994fc5e50d2090dcf209f508696551ce0f5b35a Author: Alasdair G Kergon <agk(a)redhat.com> AuthorDate: Fri Oct 6 23:10:02 2017 +0100 Committer: Alasdair G Kergon <agk(a)redhat.com> CommitterDate: Fri Oct 6 23:10:02 2017 +0100 man: regenerate --- man/lvconvert.8_pregen | 10 ++++++++-- man/pvscan.8_pregen | 7 +++---- 2 files changed, 11 insertions(+), 6 deletions(-) diff --git a/man/lvconvert.8_pregen b/man/lvconvert.8_pregen index 93a8b27..fb2d869 100644 --- a/man/lvconvert.8_pregen +++ b/man/lvconvert.8_pregen @@ -753,10 +753,12 @@ Merge COW snapshot LV into its origin. Replace failed PVs in a raid or mirror LV. .br -Repair a thin pool. +Repair a thin pool. +.br +Repair a cache pool. .br .P -\fBlvconvert\fP \fB--repair\fP \fILV\fP\fI_thinpool_mirror_raid\fP +\fBlvconvert\fP \fB--repair\fP \fILV\fP\fI_thinpool_cache_cachepool_mirror_raid\fP .br .RS 4 .ad l @@ -767,6 +769,10 @@ Repair a thin pool. [ \fB--usepolicies\fP ] .ad b .br +.ad l +[ \fB--poolmetadataspare\fP \fBy\fP|\fBn\fP ] +.ad b +.br [ COMMON_OPTIONS ] .RE .br diff --git a/man/pvscan.8_pregen b/man/pvscan.8_pregen index 4927fcb..ebd155f 100644 --- a/man/pvscan.8_pregen +++ b/man/pvscan.8_pregen @@ -76,8 +76,7 @@ When event-driven system services detect a new LVM device, the first step is to automatically scan and cache the metadata from the device. This is done by pvscan --cache. A second step is to automatically activate LVs that are present on the new device. This auto-activation is done by the -same pvscan --cache command when the option '-a|--activate ay' is -included. +same pvscan --cache command when the option --activate ay is included. Auto-activation of VGs or LVs can be enabled/disabled using: .br @@ -92,11 +91,11 @@ When this setting is undefined, all LVs are auto-activated (when lvm is fully integrated with the event-driven system services.) When a VG or LV is not auto-activated, traditional activation using -vgchange or lvchange -a|--activate is needed. +vgchange or lvchange --activate is needed. .IP \[bu] 2 pvscan auto-activation can be only done in combination with --cache. .IP \[bu] 2 -Auto-activation is designated by the "a" argument in '-a|--activate ay'. +Auto-activation is designated by the "a" argument in --activate ay. This is meant to distinguish system generated commands from explicit user commands, although it can be used in any activation command. Whenever it is used, the auto_activation_volume_list is applied.

1 0

v2_02_175 annotated tag has been created
by Alasdair Kergon 06 Oct '17

06 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=79177d1c89d99596d612e… Commit: 79177d1c89d99596d612e79e2b689336dd57cd26 Parent: 0000000000000000000000000000000000000000 Author: Alasdair G Kergon <agk(a)redhat.com> AuthorDate: 2017-10-06 21:12 +0000 Committer: Alasdair G Kergon <agk(a)redhat.com> CommitterDate: 2017-10-06 21:12 +0000 annotated tag: v2_02_175 has been created at 79177d1c89d99596d612e79e2b689336dd57cd26 (tag) tagging 73b72b8331949a1af778ad34491de3c7663c107f (commit) replaces v2_02_174 Release 2.02.175. A small release with an assortment of fixes. Also note that we do not intend to develop lvm2app or the python bindings any further, and would like any users to move towards the D-Bus interface. (Reference: https://github.com/lvmteam/lvm2/issues/1 ) 48 files changed, 712 insertions(+), 194 deletions(-) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.14 (GNU/Linux) iEYEABECAAYFAlnX8qcACgkQIoGRwVZ+LBfDMwCfVMlyP3W3510chWCuDHzCjmED McoAni9FgeEsWxBxupz3b/G7/1dCRLjs =KwpA -----END PGP SIGNATURE----- Alasdair G Kergon (5): kernel: Refresh dm-ioctl.h dmsetup: Process -S to filter input device lists. vgsplit: Fix intermediate metadata corruption. vgmerge: Fix intermediate metadata corruption pre-release David Teigland (7): revert tidy: prefer not using else after return lvcreate: use cmd defs to deny unspported lockd cases man pvscan: just write --activate in description toollib: fix parentheses in assignment + comparison lvmlockd: activate mirror LVs in shared mode with cmirrord pvmove: require LV name in a shared VG improve error messages when command rules fail Heinz Mauelshagen (1): raid: enhance conversion rejection message Marian Csontos (5): liblvm: Deprecate library and bindings liblvm: Add deprecation warning to configure configure: Update configure configure: Add warning for D-Bus w/o notifications pvmove: Change error message Peter Rajnoha (6): blkdeactivate: fix --{dm,lvm,mpath}options option name recognition blkdeactivate: fix blkdeactivate regression with failing DM/MD devs deactivation blkdeactivate: add blkdeactivate -r wait option to wait for MD resync/recovery/reshape blk-availability: use blkdeactivate -r wait in blk-availability systemd service/initscript blkdeactivate: also try to unmount /boot on blkdeactivate -u if on top of supported device blkdeactivate: cleanup Thomas Lamprecht (1): dmeventd: ensure systemd service gets stopped on shutdown Tony Asleson (6): lvmdbusd: Fix hang in MThreadRunner lvmdbusd: Ensure vg_uuid is present lvmdbusd: Improve args comparison lvmdbusd: Main thread exception logging lvmdbusd: Give threads names lvmdbusd: thread stacks dump support Zdenek Kabelac (8): lvconvert: repair supports poometadatespare switch lvconvert: support repair of cache/cachepool libdm: fix parentheses in assignment + comparison tests: check for free space in test dir tests: repair of cachepool tests: use portable df output dmeventd: schedule exit on break fsadm: add --help

1 0

master - post-release
by Alasdair Kergon 06 Oct '17

06 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=a994fc5e50d2090dcf209… Commit: a994fc5e50d2090dcf209f508696551ce0f5b35a Parent: 73b72b8331949a1af778ad34491de3c7663c107f Author: Alasdair G Kergon <agk(a)redhat.com> AuthorDate: Fri Oct 6 22:16:58 2017 +0100 Committer: Alasdair G Kergon <agk(a)redhat.com> CommitterDate: Fri Oct 6 22:16:58 2017 +0100 post-release --- VERSION | 2 +- VERSION_DM | 2 +- WHATS_NEW | 3 +++ WHATS_NEW_DM | 3 +++ 4 files changed, 8 insertions(+), 2 deletions(-) diff --git a/VERSION b/VERSION index 9611e66..e6f3580 100644 --- a/VERSION +++ b/VERSION @@ -1 +1 @@ -2.02.175(2)-git (2017-10-06) +2.02.176(2)-git (2017-10-06) diff --git a/VERSION_DM b/VERSION_DM index c5894b6..cb4da47 100644 --- a/VERSION_DM +++ b/VERSION_DM @@ -1 +1 @@ -1.02.144-git (2017-10-06) +1.02.145-git (2017-10-06) diff --git a/WHATS_NEW b/WHATS_NEW index ded5335..7899365 100644 --- a/WHATS_NEW +++ b/WHATS_NEW @@ -1,3 +1,6 @@ +Version 2.02.176 - +=================================== + Version 2.02.175 - 6th October 2017 =================================== Use --help with blockdev when checking for --getsize64 support in fsadm. diff --git a/WHATS_NEW_DM b/WHATS_NEW_DM index e8035d1..a5d24dd 100644 --- a/WHATS_NEW_DM +++ b/WHATS_NEW_DM @@ -1,3 +1,6 @@ +Version 1.02.145 - +=================================== + Version 1.02.144 - 6th October 2017 =================================== Schedule exit when received SIGTERM in dmeventd.

1 0

master - pre-release
by Alasdair Kergon 06 Oct '17

06 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=73b72b8331949a1af778a… Commit: 73b72b8331949a1af778ad34491de3c7663c107f Parent: 486ed108481cfbe4336f0fa590cbae251188ecc6 Author: Alasdair G Kergon <agk(a)redhat.com> AuthorDate: Fri Oct 6 22:10:48 2017 +0100 Committer: Alasdair G Kergon <agk(a)redhat.com> CommitterDate: Fri Oct 6 22:10:48 2017 +0100 pre-release --- VERSION | 2 +- VERSION_DM | 2 +- WHATS_NEW | 9 +++++++-- WHATS_NEW_DM | 4 ++-- 4 files changed, 11 insertions(+), 6 deletions(-) diff --git a/VERSION b/VERSION index eda932e..9611e66 100644 --- a/VERSION +++ b/VERSION @@ -1 +1 @@ -2.02.175(2)-git (2017-09-13) +2.02.175(2)-git (2017-10-06) diff --git a/VERSION_DM b/VERSION_DM index 8a32c45..c5894b6 100644 --- a/VERSION_DM +++ b/VERSION_DM @@ -1 +1 @@ -1.02.144-git (2017-09-13) +1.02.144-git (2017-10-06) diff --git a/WHATS_NEW b/WHATS_NEW index bba813d..ded5335 100644 --- a/WHATS_NEW +++ b/WHATS_NEW @@ -1,12 +1,17 @@ -Version 2.02.175 - -====================================== +Version 2.02.175 - 6th October 2017 +=================================== Use --help with blockdev when checking for --getsize64 support in fsadm. + Dump lvmdbusd debug information with SIGUSR1. Fix metadata corruption in vgsplit and vgmerge intermediate states. Add PV_MOVED_VG PV status flag to mark PVs moving between VGs. + Fix lvmdbus hang and recognise unknown VG correctly. + Improve error messages when command rules fail. Require LV name with pvmove in a shared VG. Allow shared active mirror LVs with lvmlockd, dlm, and cmirrord. Support lvconvert --repair with cache and cachepool volumes. lvconvert --repair respects --poolmetadataspare option. + Mark that we don't plan to develop liblvm2app and python bindings any further. + Fix thin pool creation in shared VG. (2.02.173) Version 2.02.174 - 13th September 2017 ====================================== diff --git a/WHATS_NEW_DM b/WHATS_NEW_DM index ea01089..e8035d1 100644 --- a/WHATS_NEW_DM +++ b/WHATS_NEW_DM @@ -1,5 +1,5 @@ -Version 1.02.144 - -====================================== +Version 1.02.144 - 6th October 2017 +=================================== Schedule exit when received SIGTERM in dmeventd. Also try to unmount /boot on blkdeactivate -u if on top of supported device. Use blkdeactivate -r wait in blk-availability systemd service/initscript.

1 0

master - vgmerge: Fix intermediate metadata corruption
by Alasdair Kergon 05 Oct '17

05 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=486ed108481cfbe4336f0… Commit: 486ed108481cfbe4336f0fa590cbae251188ecc6 Parent: a781b1c1788461595f382918bb1fc210d248d444 Author: Alasdair G Kergon <agk(a)redhat.com> AuthorDate: Fri Oct 6 02:12:42 2017 +0100 Committer: Alasdair G Kergon <agk(a)redhat.com> CommitterDate: Fri Oct 6 02:20:45 2017 +0100 vgmerge: Fix intermediate metadata corruption vgmerge suffers from a similar problem to the one fixed in commit 8146548d25e9104f0d530d943290d448c1994c0a ("vgsplit: Fix intermediate metadata corruption.") When merging, splitting or renaming VGs, use a new PV status flag PV_MOVED_VG to mark the PVs that hold metadata with the old VG name and use this to provide PV-level granularity instead of incorrectly assuming all PVs in the VG are the same. --- WHATS_NEW | 3 ++- lib/format_text/flags.c | 1 + lib/format_text/format-text.c | 15 +++++++++------ lib/metadata/metadata-exported.h | 3 ++- lib/metadata/metadata.c | 6 ++++++ tools/vgmerge.c | 5 +++++ 6 files changed, 25 insertions(+), 8 deletions(-) diff --git a/WHATS_NEW b/WHATS_NEW index 0cdd398..bba813d 100644 --- a/WHATS_NEW +++ b/WHATS_NEW @@ -1,7 +1,8 @@ Version 2.02.175 - ====================================== Use --help with blockdev when checking for --getsize64 support in fsadm. - Fix metadata corruption in vgsplit intermediate state. + Fix metadata corruption in vgsplit and vgmerge intermediate states. + Add PV_MOVED_VG PV status flag to mark PVs moving between VGs. Require LV name with pvmove in a shared VG. Allow shared active mirror LVs with lvmlockd, dlm, and cmirrord. Support lvconvert --repair with cache and cachepool volumes. diff --git a/lib/format_text/flags.c b/lib/format_text/flags.c index 156d542..51acc46 100644 --- a/lib/format_text/flags.c +++ b/lib/format_text/flags.c @@ -48,6 +48,7 @@ static const struct flag _pv_flags[] = { {EXPORTED_VG, "EXPORTED", STATUS_FLAG}, {MISSING_PV, "MISSING", COMPATIBLE_FLAG}, {MISSING_PV, "MISSING", STATUS_FLAG}, + {PV_MOVED_VG, NULL, 0}, {UNLABELLED_PV, NULL, 0}, {0, NULL, 0} }; diff --git a/lib/format_text/format-text.c b/lib/format_text/format-text.c index c359c8a..3e1ca19 100644 --- a/lib/format_text/format-text.c +++ b/lib/format_text/format-text.c @@ -609,14 +609,17 @@ static int _vg_write_raw(struct format_instance *fid, struct volume_group *vg, struct mda_header *mdah; struct pv_list *pvl; int r = 0; - uint64_t new_wrap = 0, old_wrap = 0, new_end; + uint64_t new_wrap = 0, old_wrap = 0, new_end; int found = 0; int noprecommit = 0; + const char *old_vg_name = NULL; /* Ignore any mda on a PV outside the VG. vgsplit relies on this */ dm_list_iterate_items(pvl, &vg->pvs) { if (pvl->pv->dev == mdac->area.dev) { found = 1; + if (pvl->pv->status & PV_MOVED_VG) + old_vg_name = vg->old_name; break; } } @@ -630,8 +633,7 @@ static int _vg_write_raw(struct format_instance *fid, struct volume_group *vg, if (!(mdah = raw_read_mda_header(fid->fmt, &mdac->area))) goto_out; - rlocn = _find_vg_rlocn(&mdac->area, mdah, - vg->old_name ? vg->old_name : vg->name, &noprecommit); + rlocn = _find_vg_rlocn(&mdac->area, mdah, old_vg_name ? : vg->name, &noprecommit); mdac->rlocn.offset = _next_rlocn_offset(rlocn, mdah); if (!fidtc->raw_metadata_buf && @@ -719,11 +721,14 @@ static int _vg_commit_raw_rlocn(struct format_instance *fid, int r = 0; int found = 0; int noprecommit = 0; + const char *old_vg_name = NULL; /* Ignore any mda on a PV outside the VG. vgsplit relies on this */ dm_list_iterate_items(pvl, &vg->pvs) { if (pvl->pv->dev == mdac->area.dev) { found = 1; + if (pvl->pv->status & PV_MOVED_VG) + old_vg_name = vg->old_name; break; } } @@ -734,9 +739,7 @@ static int _vg_commit_raw_rlocn(struct format_instance *fid, if (!(mdah = raw_read_mda_header(fid->fmt, &mdac->area))) goto_out; - if (!(rlocn = _find_vg_rlocn(&mdac->area, mdah, - vg->old_name ? vg->old_name : vg->name, - &noprecommit))) { + if (!(rlocn = _find_vg_rlocn(&mdac->area, mdah, old_vg_name ? : vg->name, &noprecommit))) { mdah->raw_locns[0].offset = 0; mdah->raw_locns[0].size = 0; mdah->raw_locns[0].checksum = 0; diff --git a/lib/metadata/metadata-exported.h b/lib/metadata/metadata-exported.h index 92b38de..4a8c519 100644 --- a/lib/metadata/metadata-exported.h +++ b/lib/metadata/metadata-exported.h @@ -84,6 +84,7 @@ #define CONVERTING UINT64_C(0x0000000000400000) /* LV */ #define MISSING_PV UINT64_C(0x0000000000800000) /* PV */ +#define PV_MOVED_VG UINT64_C(0x4000000000000000) /* PV - Moved to a new VG */ #define PARTIAL_LV UINT64_C(0x0000000001000000) /* LV - derived flag, not written out in metadata*/ @@ -146,7 +147,7 @@ #define LV_RESHAPE UINT64_C(0x1000000000000000) /* Ongoing reshape (number of stripes, stripesize or raid algorithm change): used as SEGTYPE_FLAG to prevent activation on old runtime */ #define LV_RESHAPE_DATA_OFFSET UINT64_C(0x2000000000000000) /* LV reshape flag data offset (out of place reshaping) */ -/* Next unused flag: UINT64_C(0x4000000000000000) */ +/* Next unused flag: UINT64_C(0x8000000000000000) */ /* Format features flags */ #define FMT_SEGMENTS 0x00000001U /* Arbitrary segment params? */ diff --git a/lib/metadata/metadata.c b/lib/metadata/metadata.c index bbbf7d0..c50d579 100644 --- a/lib/metadata/metadata.c +++ b/lib/metadata/metadata.c @@ -520,6 +520,8 @@ int vg_rename(struct cmd_context *cmd, struct volume_group *vg, pv_dev_name(pvl->pv)); return 0; } + /* Mark the PVs that still hold metadata with the old VG name */ + pvl->pv->status |= PV_MOVED_VG; } return 1; @@ -3616,6 +3618,7 @@ static int _vg_commit_mdas(struct volume_group *vg) int vg_commit(struct volume_group *vg) { int cache_updated = 0; + struct pv_list *pvl; if (!lvmcache_vgname_is_locked(vg->name)) { log_error(INTERNAL_ERROR "Attempt to write new VG metadata " @@ -3631,11 +3634,14 @@ int vg_commit(struct volume_group *vg) /* Instruct remote nodes to upgrade cached metadata. */ if (!remote_commit_cached_metadata(vg)) stack; // FIXME: What should we do? + /* * We need to clear old_name after a successful commit. * The volume_group structure could be reused later. */ vg->old_name = NULL; + dm_list_iterate_items(pvl, &vg->pvs) + pvl->pv->status &= ~PV_MOVED_VG; /* This *is* the original now that it's commited. */ release_vg(vg->vg_committed); diff --git a/tools/vgmerge.c b/tools/vgmerge.c index d49e4e1..86c6a48 100644 --- a/tools/vgmerge.c +++ b/tools/vgmerge.c @@ -111,6 +111,8 @@ static int _vgmerge_single(struct cmd_context *cmd, const char *vg_name_to, del_pvl_from_vgs(vg_from, pvl); add_pvl_to_vgs(vg_to, pvl); pvl->pv->vg_name = dm_pool_strdup(cmd->mem, vg_to->name); + /* Mark the VGs that still hold metadata for the old VG */ + pvl->pv->status |= PV_MOVED_VG; } /* Fix up LVIDs */ @@ -168,6 +170,9 @@ static int _vgmerge_single(struct cmd_context *cmd, const char *vg_name_to, vg_to->extent_count += vg_from->extent_count; vg_to->free_count += vg_from->free_count; + /* Flag up that some PVs have moved from another VG */ + vg_to->old_name = vg_from->name; + /* store it on disks */ log_verbose("Writing out updated volume group"); if (!vg_write(vg_to) || !vg_commit(vg_to))

1 0

master - dmeventd: ensure systemd service gets stopped on shutdown
by Peter Rajnoha 05 Oct '17

05 Oct '17

Gitweb: https://sourceware.org/git/?p=lvm2.git;a=commitdiff;h=a781b1c1788461595f382… Commit: a781b1c1788461595f382918bb1fc210d248d444 Parent: 3ae8adce928d4ce6ef36c87c2a0bb8aa737346b8 Author: Thomas Lamprecht <t.lamprecht(a)proxmox.com> AuthorDate: Thu Oct 5 12:52:15 2017 +0200 Committer: Peter Rajnoha <prajnoha(a)redhat.com> CommitterDate: Thu Oct 5 13:01:55 2017 +0200 dmeventd: ensure systemd service gets stopped on shutdown Add these for dmeventd systemd unit (dm-event.service): Before: shutdown.target Conflicts: shutdown.target This will cause the dmeventd to be properly stopped at shutdown (after all the dmeventd clients unregistered their devices from monitoring) with dm-event.service's stop action (there's no direct action defined for the "stop" so systemd sends SIGTERM instead). Before, we let dmeventd to get killed only as part of the very last SIGTERM/SIGKILL for all the remaining processes late in the shutdown sequence so we may have missed some logs if dmeventd encountered an error during its shutdown (logging facilities are already off at this late time in shutdown sequence). Ref: https://www.redhat.com/archives/lvm-devel/2017-October/msg00000.html --- scripts/dm_event_systemd_red_hat.service.in | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/scripts/dm_event_systemd_red_hat.service.in b/scripts/dm_event_systemd_red_hat.service.in index 7c607aa..4924d39 100644 --- a/scripts/dm_event_systemd_red_hat.service.in +++ b/scripts/dm_event_systemd_red_hat.service.in @@ -3,7 +3,8 @@ Description=Device-mapper event daemon Documentation=man:dmeventd(8) Requires=dm-event.socket After=dm-event.socket -Before=local-fs-pre.target +Before=local-fs-pre.target shutdown.target +Conflicts=shutdown.target DefaultDependencies=no [Service]

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

lvm2-commits October 2017