Skip to contentSkip to conversation

pvs complains of missing PVs that are not missing

linux-lvm Joe Harvell 2 participants 7 messages Oct 9, 2013 – Oct 10, 2013 3 messages not in the archive

lore
Share & tools
lore
#0 lore raw

As part of moving disks from one system to another, somehow LVM thinks PVS are missing in the new system but they are not. Before I explain the relevant chain of events related to moving the disks, here is the problem as it manifests in the new system:

joey@akita ~ $ sudo pvs -v
Scanning for physical volume names
There are 9 physical volumes missing.
There are 9 physical volumes missing.
There are 9 physical volumes missing.
PV VG Fmt Attr PSize PFree DevSize PV UUID
/dev/md/saluki:r10_0 bayeux lvm2 a-m 40,00g 0 40,00g
C4V6YY-1bA1-Clg9-f45q-ffma-BH07-K7ex1w
/dev/md/saluki:r10_1 bayeux lvm2 --m 74,97g 0 75,00g
d3LmqC-1GnU-LPjb-uYqH-Z2jG-lQrp-31JOXo
/dev/md/saluki:r10_2 bayeux lvm2 a-m 224,78g 154,78g 224,81g
JQsXS2-XfhA-zucx-Xetf-HHI0-aGFp-Ibz2We
/dev/md/saluki:r5_0 bayeux lvm2 a-m 360,00g 2,00g 360,00g
UcdCOb-RqEK-0ofL-ql90-viLC-ssri-eopnTg
/dev/md/saluki:r5_1 bayeux lvm2 a-m 279,09g 145,09g 279,11g
FLFUQW-PZHK-uY19-Zblf-s3RU-QPZK-0Pc3vC
/dev/sda11 bayeux lvm2 a-- 93,31g 93,31g 93,32g
KSVeZ5-DUI8-XCK4-NfiF-WB2r-6RGe-GymTuF
/dev/sda5 bayeux lvm2 a-- 150,00g 0 150,00g
FhGyS2-yKGw-pfxE-EyY4-yGi3-3Hoa-JCRCk1
/dev/sda7 bayeux lvm2 a-- 107,16g 59,59g 107,16g
lghpEG-bnje-tBY3-1jGJ-suAN-S8g5-ti5Df0
/dev/sda9 bayeux lvm2 a-- 29,38g 22,34g 29,38g
8wXKU8-2phP-4NKE-hVMo-8VY2-5Z7D-SVRwmU
/dev/sdb3 seward lvm2 a-- 234,28g 104,28g 234,28g
WnNkO0-8709-p5lN-bTGF-KdAJ-X29B-1cM5bv
/dev/sdc11 bayeux lvm2 a-- 93,31g 86,28g 93,32g
MoWrvQ-oI3A-OWBT-cwkp-eswH-BkNp-fuhXLI
/dev/sdc5 bayeux lvm2 a-- 150,00g 150,00g 150,00g
eeVLsy-DIb3-1w1G-VtIa-S6Bv-w9Li-pVQhLD
/dev/sdc7 bayeux lvm2 a-- 107,16g 107,16g 107,16g
K8ibVQ-AABO-islF-imv0-a0wv-ho4w-mxAUBO
/dev/sdc9 bayeux lvm2 a-- 29,38g 29,38g 29,38g
csjMOF-pIO8-o2dP-Vm5l-QRhP-6g5G-UdOSqH
/dev/sdd1 shanghai lvm2 --- 2,73t 0 2,73t
IVJKal-Oode-Yn0T-oS9z-tadX-X1cs-1J2ut1
/dev/sde1 bayeux lvm2 a-m 372,59g 172,59g 372,61g
e21dTH-FxZw-P4Ug-f0S1-jIe9-hdYc-MCyH1D
/dev/sdf11 bayeux lvm2 a-m 93,31g 76,88g 93,32g
jPQ6YS-LPTg-N7jA-65CA-F0tP-VSSE-GZ6vX5
/dev/sdf6 bayeux lvm2 a-m 150,00g 0 150,00g
cxcme3-i1US-6MZI-Nx2U-fkSg-gQfz-X0Y468
/dev/sdf7 bayeux lvm2 a-- 107,16g 107,16g 107,16g
pTZN6n-sLvt-whyf-rWkJ-bZoP-IkiE-lEe92P
/dev/sdf9 bayeux lvm2 a-m 29,38g 128,00m 29,38g
Qtqo8m-UjUh-Qe4R-VkkN-oJmK-pHG3-lPJVSm

joey@akita ~ $ sudo pvscan

  PV /dev/sdd1              VG shanghai   lvm2 [2,73 TiB / 0 free]
  PV /dev/sdb3              VG seward     lvm2 [234,28 GiB / 104,28 GiB 
free]
  PV /dev/md/saluki:r10_0   VG bayeux     lvm2 [40,00 GiB / 0 free]
  PV /dev/md/saluki:r5_0    VG bayeux     lvm2 [360,00 GiB / 2,00 GiB free]
  PV /dev/sdf6              VG bayeux     lvm2 [150,00 GiB / 0 free]
  PV /dev/sdf7              VG bayeux     lvm2 [107,16 GiB / 107,16 GiB 
free]
  PV /dev/md/saluki:r10_1   VG bayeux     lvm2 [74,97 GiB / 0 free]
  PV /dev/sdf9              VG bayeux     lvm2 [29,38 GiB / 128,00 MiB 

free]
PV /dev/sdc5 VG bayeux lvm2 [150,00 GiB / 150,00 GiB
free]
PV /dev/sdc7 VG bayeux lvm2 [107,16 GiB / 107,16 GiB
free]

  PV /dev/sdc9              VG bayeux     lvm2 [29,38 GiB / 29,38 GiB free]
  PV /dev/sda5              VG bayeux     lvm2 [150,00 GiB / 0 free]
  PV /dev/sda7              VG bayeux     lvm2 [107,16 GiB / 59,59 GiB 
free]
  PV /dev/sda9              VG bayeux     lvm2 [29,38 GiB / 22,34 GiB free]
  PV /dev/sda11             VG bayeux     lvm2 [93,31 GiB / 93,31 GiB free]
  PV /dev/sdc11             VG bayeux     lvm2 [93,31 GiB / 86,28 GiB free]
  PV /dev/sdf11             VG bayeux     lvm2 [93,31 GiB / 76,88 GiB free]
  PV /dev/sde1              VG bayeux     lvm2 [372,59 GiB / 172,59 GiB 

free]
PV /dev/md/saluki:r5_1 VG bayeux lvm2 [279,09 GiB / 145,09 GiB
free]
PV /dev/md/saluki:r10_2 VG bayeux lvm2 [224,78 GiB / 154,78 GiB
free]
Total: 20 [5,39 TiB] / in use: 20 [5,39 TiB] / in no VG: 0 [0 ]

The problem is that each of the PVs that show up with the "missing" attribute are actually present in the new system.

The way I would like to prove that they are in the system is to read the PV disk label directly from the devices and show they have the same UUID. But I am unable to find how to do that. So instead I'll take a simple example from the above 9 missing PVs. pvs shows that the PV with UUID e21dTH-FxZw-P4Ug-f0S1-jIe9-hdYc-MCyH1D should be found at /dev/sde1 but it is not found in the system. I know for a fact that it is in the system. Here's how I know.

I know that the Seagate disk (the only Seagate in the system) has 1 partition that is a PV in VG bayeux, and that it contains exactly one LV named backup. Here's what the system shows now about /dev/sde1:

joey@akita /tmp $ sudo parted /dev/sde print Mot de passe :
Mod�le: ATA ST3400620AS (scsi)
Disque /dev/sde : 400GB
Taille des secteurs (logiques/physiques): 512B/512B Table de partitions : gpt
Disk Flags:

Num�ro D�but Fin Taille Syst�me de fichiers Nom Fanions
1 17,4kB 400GB 400GB bayeux lvm
(gestionnaire de volumes logiques)

So /dev/sde1 is the place where /dev/bayeux/backup should reside.

Here is what the system shows now about that LV and the (supposedly missing) /dev/sde1 PV:

joey@akita /tmp $ sudo pvdisplay --maps /dev/sde1
--- Physical volume ---

  PV Name               /dev/sde1
  VG Name               bayeux
  PV Size               372,61 GiB / not usable 18,05 MiB
  Allocatable           yes
  PE Size               32,00 MiB
  Total PE              11923
  Free PE               5523
  Allocated PE          6400
  PV UUID               e21dTH-FxZw-P4Ug-f0S1-jIe9-hdYc-MCyH1D

--- Physical Segments ---
Physical extent 0 to 6399:
Logical volume /dev/bayeux/backup
Logical extents 0 to 6399
Physical extent 6400 to 11922:
FREE

joey@akita /tmp $ sudo lvdisplay --maps /dev/bayeux/backup
--- Logical volume ---

  LV Path                /dev/bayeux/backup
  LV Name                backup
  VG Name                bayeux
  LV UUID                QhTdFi-cGuL-380h-2hIB-wXNx-D57w-pdlPpB
  LV Write Access        read/write
  LV Creation host, time saluki, 2013-05-11 17:14:45 -0500
  LV Status              available
  # open                 0
  LV Size                200,00 GiB
  Current LE             6400
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     256
  Block device           254:32

--- Segments ---
Logical extent 0 to 6399:

    Type                linear
    Physical volume     /dev/sde1
    Physical extents    0 to 6399

And to top it off, I can mount the filesystem that lives on /dev/sde1 and I see the backup data (I activated bayeux with --partial):

joey@akita /tmp $ sudo mkdir -m 000 /tmp/bb joey@akita /tmp $ sudo mount -o ro /dev/bayeux/backup /tmp/bb joey@akita /tmp $ sudo find /tmp/bb -maxdepth 3 -ls

     2    4 drwxr-xr-x   4 root     root         4096 mai 12 00:31 /tmp/bb
7585793    4 drwxr-xr-x   4 root     root         4096 mai 12 00:33 /tmp/bb/to
7585794    4 drwx------   2 root     root         4096 nov. 27 2012 /tmp/bb/to/lost+found
7585795    4 drwxr-xr-x   3 marta    marta        4096 mai 12 01:11 /tmp/bb/to/pikawa
262145    8 -rw-r--r--   1 marta    marta        6148 d�c. 10 2012 /tmp/bb/to/pikawa/.DS_Store
7585796    4 drwx--S---   3 marta    marta        4096 mai 12 01:11 /tmp/bb/to/pikawa/pikawa.sparsebundle
    11   16 drwx------   2 root     root        16384 mai 11 23:35 

/tmp/bb/lost+found

Ok, I think that shows the PV is in fact not missing from the system! Now here's the explanation of the chain of events that I think contributed to getting in this state:

1.  Initially, I had 5 disks in host saluki.  Three Western Digital 1TB disks, 1 Western Digital 3TB disk, and 1 Seagate 400 GB disk.
2.  I removed all but two of the 1TB WD disks from saluki and rebooted it.  I was able to boot due to a combination of Linux raid (not LVM RAID)  and non-essential file systems on the disks I removed.
3.  Then I re-added all the disks to saluki except the Seagate.  I re-added the partitions on the WD 1TB I had removed to the corresponding RAID volumes, and all the RAID volumes re-synced fine.  I re-added all the file systems to /etc/fstab except the one on the Seagate and continued in that configuration for a while.
4.  Finally, I moved all 5 disks into host akita.  akita is a new machine and I installed the entire OS in a new VG (seward) consisting of 1 disk.

I think the 9 PVs that are having the problems are related to the disks I removed and later re-added. I should point out that between steps 3 and 4 I tried to do an lvextend in VG bayeux, but it told me it would not allow this since it had partial PVs. Of course that makes total sense, and I'm glad it stopped me. So that means I cold not have changed any LVM configuration for bayeux while the VG was incomplete. So I don't see any reason why LVM is complaining about missing PVs.

FYI, here are the details of the RAID configuration:

joey@akita /tmp $ cat /proc/mdstat Personalities : [raid1] [raid10] [raid6] [raid5] [raid4] md121 : active raid10 sda6[0] sdf5[2] sdc6[1]

      235732992 blocks super 1.2 512K chunks 2 near-copies [3/3] [UUU]
      bitmap: 0/2 pages [0KB], 65536KB chunk

md122 : active raid1 sdc1[4] sda1[5] sdf1[1]

      102436 blocks super 1.0 [3/3] [UUU]
      bitmap: 0/7 pages [0KB], 8KB chunk

md123 : active raid10 sdc2[0] sda2[3] sdf2[1]

      8388864 blocks super 1.0 64K chunks 2 near-copies [3/3] [UUU]
      bitmap: 0/9 pages [0KB], 512KB chunk

md124 : active raid10 sdc3[4] sda3[5] sdf3[1]

      41943232 blocks super 1.0 64K chunks 2 near-copies [3/3] [UUU]
      bitmap: 0/161 pages [0KB], 128KB chunk

md125 : active raid5 sdc4[4] sda4[5] sdf4[3]

      377487616 blocks super 1.0 level 5, 128k chunk, algorithm 2 [3/3] 
[UUU]
      bitmap: 0/181 pages [0KB], 512KB chunk

md126 : active raid10 sda8[4] sdf8[2] sdc8[3]

      78643008 blocks super 1.0 64K chunks 2 near-copies [3/3] [UUU]
      bitmap: 0/1 pages [0KB], 65536KB chunk

md127 : active raid5 sda10[0] sdf10[3] sdc10[1]

      292663040 blocks super 1.0 level 5, 128k chunk, algorithm 2 [3/3] 
[UUU]
      bitmap: 0/140 pages [0KB], 512KB chunk

unused devices: <none>
joey@akita /tmp $ ls -l /dev/md/*

lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:boot -> ../md122
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:r10_0 -> ../md124
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:r10_1 -> ../md126
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:r10_2 -> ../md121
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:r5_0 -> ../md125
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:r5_1 -> ../md127
lrwxrwxrwx 1 root root 8  8 oct.  23:51 /dev/md/saluki:swap -> ../md123

Note that all of the RAID volumes are on partitions of disks sda, sdc and sdf (the three 1 TB WDs):

joey@akita /dev/disk/by-id $ ls -l ata* | egrep 'sd[acf]$'

lrwxrwxrwx 1 root root  9  8 oct.  23:51 ata-WDC_WD1001FALS-00E8B0_WD-WMATV6936241 -> ../../sdc
lrwxrwxrwx 1 root root  9  9 oct.  15:24 ata-WDC_WD1001FALS-00J7B0_WD-WMATV0666975 -> ../../sdf
lrwxrwxrwx 1 root root  9  8 oct.  23:51 ata-WDC_WD1001FALS-00J7B0_WD-WMATV6998349 -> ../../sda
#1 lore raw
Quoted 4 lines

As part of moving disks from one system to another, somehow LVM thinks PVS are missing in the new system but they are not. Before I explain the relevant chain of events related to moving the disks, here is the problem as it manifests in the new system:

[snip]

I forgot to mention the LVM version info:

joey@akita /dev/disk/by-id $ sudo lvm version
LVM version: 2.02.103(2) (2013-10-04)
Library version: 1.02.82 (2013-10-04)
Driver version: 4.25.0
j

A message that is not in the archive (sent off-list or not yet received): [email protected]

#3 lore raw

Le 09/10/2013 17:09, matthew patton a �crit :

Quoted 2 lines

clear LVM's cache (or just disable it in lvm.conf), rescan pv, vg, lv. make sure udev is creating the necessary hooks.

I'm not sure how to disable the cache. Here are my relevant settings in /etc/lvm/lvm.conf:

# If set, the cache of block device nodes with all associated symlinks # will be constructed out of the existing udev database content. # This avoids using and opening any inapplicable non-block devices or # subdirectories found in the device directory. This setting is applied # to udev-managed device directory only, other directories will be
scanned

    # fully. LVM2 needs to be compiled with udev support for this 
setting to
    # take effect. N.B. Any device node or symlink not managed by udev in
    # udev directory will be ignored with this setting on.
    obtain_device_list_from_udev = 1

# The results of the filtering are cached on disk to avoid # rescanning dud devices (which can take a very long time). # By default this cache is stored in the /etc/lvm/cache directory # in a file called '.cache'.
# It is safe to delete the contents: the tools regenerate it. # (The old setting 'cache' is still respected if neither of # these new ones is present.)
# N.B. If obtain_device_list_from_udev is set to 1 the list of # devices is instead obtained from udev and any existing .cache # file is removed.
cache_dir = "/etc/lvm/cache"
cache_file_prefix = ""

# You can turn off writing this cache file by setting this to 0. write_cache_state = 1

And as expected based on my setting for obtain_device_list_from_udev, there is no /etc/lvm/cache directorypresent:

joey@akita ~ $ sudo nano -w /etc/lvm/lvm.conf Mot de passe :
joey@akita ~ $ ls -al /etc/lvm/
total 68

drwxr-xr-x  5 root root  4096  7 oct.  14:12 .
drwxr-xr-x 74 root root  4096  9 oct.  11:21 ..
drwx------  2 root root  4096  8 oct.  22:21 archive
drwx------  2 root root  4096  8 oct.  22:21 backup
-rw-r--r--  1 root root 45697  9 oct.  17:27 lvm.conf
drwxr-xr-x  2 root root  4096  8 oct.  19:24 profile

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#6 lore raw

Le 09/10/2013 18:30, matthew patton a �crit :

Quoted 5 lines
Quoted 2 lines

So I should set 'obtain_device_list_from_udev' to 0, then pvscan, vgscan and lvscan?

worth a shot. have you confirmed that udev has all the basic disk devices created?

I tried that to no avail. Yes, all the block devices were present in /dev, both for the raw partitions and the RAID ones.

Does anyone know the algorithm LVM uses to determine whether PVs are present? Also, I'd really like an LVM tool that reads the PV label off of a PV and displays it...I want to see what UUID label is actually on each PV.

#7 lore raw
Quoted 7 lines · Joe Harvell

Le 09/10/2013 18:30, matthew patton a �crit :

Quoted 6 lines
Quoted 2 lines

So I should set 'obtain_device_list_from_udev' to 0, then pvscan, vgscan and lvscan?

worth a shot. have you confirmed that udev has all the basic disk devices created?

Quoted 8 lines

I tried that to no avail. Yes, all the block devices were present in /dev, both for the raw partitions and the RAID ones.

Does anyone know the algorithm LVM uses to determine whether PVs are present? Also, I'd really like an LVM tool that reads the PV label off of a PV and displays it...I want to see what UUID label is actually on each PV.

Two important questions here - which distro is this?

There are two notions of cache in LVM. One is device cache, the other one is metadata cache. The first one is controlled by write_cache_state setting (which is obsoleted by obtaining the device list from udev). The latter one is controlled by use_lvmetad setting. The lvmetad (and metadata cache) has been added to LVM just recently, while the device cache is there for a long time...

As for the other important question:
Is lvmetad used or not? (check global/use_lvmetad lvm.conf setting). If lvmetad is used, it gathers incoming PVs based on events which means once the PV is available in the system, lvmetad gets notified automatically. Then the PV is scanned for LVM metadata and lvmetad stores that information. This information is then reused for each LVM command call instead of scanning the /dev again and again for PVs. The lvmetad requires udev for its operation! If lvmetad is used, does pvscan --cache call help?

If lvmetad is not used, whenever the LVM command is executed, each block device in /dev is scanned for PV labels, every time! Here, the obtain_device_list_from_udev lvm.conf setting makes a difference in a way how we get the list of block devices - if this setting is disabled, LVM directly scans all the /dev content and it selects block devices itself. If it's enabled, we get the list of block devices from udev database (which is a bit quicker as we don't need to iterate over all the content of /dev and decide which item is a block device or not, saving a bit of time this way).

Peter

#8 lore raw
Quoted 4 lines · Peter Rajnoha

There are two notions of cache in LVM. One is device cache, the other one is metadata cache. The first one is controlled by write_cache_state setting (which is obsoleted by obtaining the device list from udev).

...

Quoted 6 lines

If lvmetad is not used, whenever the LVM command is executed, each block device in /dev is scanned for PV labels, every time! Here, the obtain_device_list_from_udev lvm.conf setting makes a difference in a way how we get the list of block devices - if this setting is disabled, LVM directly scans all the /dev content and it selects block devices itself.

...if LVM selects block devices itself, it can cache this 'selection' by using 'write_cache_state=1' - LVM will write this list of block devices to a cache file which is read next time, hence LVM does not need to select them once again on next command execution. Of course, this cache file is of no use when 'obtain_device_list_from_udev=1' and we get that list of block devices from udev all the time... that's the reason why 'obtain_device_list_from_udev=1' obsoletes 'write_cache_state=1'. This is the *device cache* - either from udev or from file written by LVM itself.

Peter

#9 lore raw
In reply to Peter Rajnoha

Le 10/10/2013 13:38, Peter Rajnoha a �crit :

Quoted 43 lines · Joe Harvell
Quoted 6 lines

Le 09/10/2013 18:30, matthew patton a �crit :

Quoted 5 lines
Quoted 2 lines

So I should set 'obtain_device_list_from_udev' to 0, then pvscan, vgscan and lvscan?

worth a shot. have you confirmed that udev has all the basic disk devices created?

Quoted 8 lines

I tried that to no avail. Yes, all the block devices were present in /dev, both for the raw partitions and the RAID ones.

Does anyone know the algorithm LVM uses to determine whether PVs are present? Also, I'd really like an LVM tool that reads the PV label off of a PV and displays it...I want to see what UUID label is actually on each PV.

Two important questions here - which distro is this?

There are two notions of cache in LVM. One is device cache, the other one is metadata cache. The first one is controlled by write_cache_state setting (which is obsoleted by obtaining the device list from udev). The latter one is controlled by use_lvmetad setting. The lvmetad (and metadata cache) has been added to LVM just recently, while the device cache is there for a long time...

As for the other important question:
Is lvmetad used or not? (check global/use_lvmetad lvm.conf setting). If lvmetad is used, it gathers incoming PVs based on events which means once the PV is available in the system, lvmetad gets notified automatically. Then the PV is scanned for LVM metadata and lvmetad stores that information. This information is then reused for each LVM command call instead of scanning the /dev again and again for PVs. The lvmetad requires udev for its operation! If lvmetad is used, does pvscan --cache call help?

If lvmetad is not used, whenever the LVM command is executed, each block device in /dev is scanned for PV labels, every time! Here, the obtain_device_list_from_udev lvm.conf setting makes a difference in a way how we get the list of block devices - if this setting is disabled, LVM directly scans all the /dev content and it selects block devices itself. If it's enabled, we get the list of block devices from udev database (which is a bit quicker as we don't need to iterate over all the content of /dev and decide which item is a block device or not, saving a bit of time this way).

Quoted 8 lines

Peter

_______________________________________________
linux-lvm mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/

Thanks, Peter. I run Gentoo, and I have a custom busybox based initrd with a static /dev. The initrd loads the necessary kernel modules, assembles the raid arrays and then runs lvm vgscan, followed by 'lvmchange -ay seward' before mounting the root file system and doing a switch_root exec'ing systemd. So there are two LVM installations....the initrd and the system. The lvm.conf for the initrd and for the running system are both included below with comments removed. In both cases, use_lvmetad=0 and obtain_device_list_from_udev = 1.

Is there some LVM tool that reads and display the PV label from a specified block device? Or that reads and displays the LVM meta-data from a specified block device?

Also, I realize I failed to mention something important. When I first booted with all 5 disks in akita (the new system), my static /dev in the initrd did not have all the block device entries for all the partitions of all the disks in the system. Additionally, after I added the new entries with mknod, My devices {filter} config in the initrd was filtering them out. I've since corrected both of these. Before I fixed this, I would get complaints running vgscan about several missing PVs....and the complaint output would refer to them by UUID. After I fixed this, I don't see those complaints. So that must mean it found the PVs with those UUIDs....but still it shows the PVs with the "missing" flag and forces me to use --partial when activating the vg :(

On my running system, I have already done the following procedure and saw no changes:

1. Set devices { filter = [] }
2. Set devices { obtain_device_list_from_udev = 0 }
3. Executed pvscan
4. Executed vgscan
5. Executed lvscan
6. Observed a new file created /etc/lvm/cache/.cache
7. Observed PVs still have missing flag and I still have to say --partial to activate bayeux
8. Reverted config from steps 1 and 2
9. Deleted /etc/lvm/cache and its contents

Also for reference, here is what my initrd does exactly:

#!/bin/busybox sh

export PATH=/bin:/sbin
mount -n -t proc none /proc
mount -n -t sysfs none /sys
modprobe ahci
modprobe sd_mod
modprobe dm-mod
modprobe md-mod
modprobe raid1
modprobe raid10
modprobe raid5
modprobe ext4
modprobe xhci-hcd
modprobe ehci-hcd
modprobe uhci-hcd
modprobe ohci-hcd
modprobe usbhid
modprobe hid-generic
modprobe nvidia
mdadm --assemble --scan
lvm vgscan
lvm vgchange -ay seward
mount -t ext4 /dev/seward/root /mnt
umount /proc
umount /sys
exec switch_root /mnt /usr/lib/systemd/systemd --log-level=debug --log-target=journal

// Initrd LVM version (statically linked)
LVM version: 2.02.103(2) (2013-10-04)
Library version: 1.02.82 (2013-10-04)
Driver version: 4.25.0
// Initrd LVM config

config {
checks = 1
abort_on_errors = 0
profile_dir = "/etc/lvm/profile"
}

devices {
    dir = "/dev"
    scan = [ "/dev" ]
    obtain_device_list_from_udev = 1
    preferred_names = [ ]
    filter = [ "a|^/dev/md/.*|", "a|^/dev/sd[a-f][1-9]$|", 

"a|^/dev/sd[a-f]1[0-5]$|", "r|.*|" ]
cache_dir = "/etc/lvm/cache"
cache_file_prefix = ""
write_cache_state = 1
sysfs_scan = 1
multipath_component_detection = 1
md_component_detection = 1
md_chunk_alignment = 1
data_alignment_detection = 1
data_alignment = 0
data_alignment_offset_detection = 1
ignore_suspended_devices = 0
disable_after_error_count = 0
require_restorefile_with_uuid = 1
pv_min_size = 2048
issue_discards = 0
}

allocation {
maximise_cling = 1
thin_pool_metadata_require_separate_pvs = 0
}

log {
verbose = 0
silent = 0
syslog = 1
overwrite = 0
level = 0
indent = 1
command_names = 0
prefix = " "
debug_classes = [ "memory", "devices", "activation", "allocation",
"lvmetad", "metadata", "cache", "locking" ]
}

backup {
backup = 1
backup_dir = "/etc/lvm/backup"
archive = 1
archive_dir = "/etc/lvm/archive"
retain_min = 10
retain_days = 30
}

shell {
    history_size = 100
}

global {
umask = 077
test = 0
units = "h"
si_unit_consistency = 1
activation = 1
fallback_to_lvm1 = 0
proc = "/proc"
locking_type = 1
wait_for_locks = 1
fallback_to_clustered_locking = 1
fallback_to_local_locking = 1
locking_dir = "/run/lock/lvm"
prioritise_write_locks = 1
abort_on_internal_errors = 0
detect_internal_vg_cache_corruption = 0
metadata_read_only = 0
mirror_segtype_default = "raid1"
raid10_segtype_default = "raid10"
use_lvmetad = 0
}

activation {
checks = 0
udev_sync = 1
udev_rules = 1
verify_udev_operations = 0
retry_deactivation = 1
missing_stripe_filler = "error"
use_linear_target = 1
reserved_stack = 64
reserved_memory = 8192
process_priority = -18
raid_region_size = 512
readahead = "auto"
raid_fault_policy = "warn"
mirror_log_fault_policy = "allocate"
mirror_image_fault_policy = "remove"
snapshot_autoextend_threshold = 100
snapshot_autoextend_percent = 20
thin_pool_autoextend_threshold = 100
thin_pool_autoextend_percent = 20
use_mlockall = 0
monitoring = 1
polling_interval = 15
}

metadata {
}

dmeventd {
}

dmeventd {
mirror_library = "libdevmapper-event-lvm2mirror.so" snapshot_library = "libdevmapper-event-lvm2snapshot.so" thin_library = "libdevmapper-event-lvm2thin.so"
}

// Running system LVM version:
LVM version: 2.02.103(2) (2013-10-04)
Library version: 1.02.82 (2013-10-04)
Driver version: 4.25.0

// Running system LVM config:
config {
checks = 1
abort_on_errors = 0
profile_dir = "/etc/lvm/profile"
}

devices {
    dir = "/dev"
    scan = [ "/dev" ]
    obtain_device_list_from_udev = 1
    preferred_names = [ ]
    filter = [ "a|^/dev/md/.*|", "a|^/dev/sd[a-g][1-9]$|", 

"a|^/dev/sd[a-g]1[0-5]$|", "r|.*|" ]
cache_dir = "/etc/lvm/cache"
cache_file_prefix = ""
write_cache_state = 1
sysfs_scan = 1
multipath_component_detection = 1
md_component_detection = 1
md_chunk_alignment = 1
data_alignment_detection = 1
data_alignment = 0
data_alignment_offset_detection = 1
ignore_suspended_devices = 0
disable_after_error_count = 0
require_restorefile_with_uuid = 1
pv_min_size = 2048
issue_discards = 0
}

allocation {
maximise_cling = 1
mirror_logs_require_separate_pvs = 0
thin_pool_metadata_require_separate_pvs = 0
}

log {
verbose = 0
silent = 0
syslog = 1
overwrite = 0
level = 0
indent = 1
command_names = 0
prefix = " "
debug_classes = [ "memory", "devices", "activation", "allocation",
"lvmetad", "metadata", "cache", "locking" ]
}

backup {
backup = 1
backup_dir = "/etc/lvm/backup"
archive = 1
archive_dir = "/etc/lvm/archive"
retain_min = 10
retain_days = 30
}

shell {
    history_size = 100
}

global {
umask = 077
test = 0
units = "h"
si_unit_consistency = 1
activation = 1
fallback_to_lvm1 = 0
proc = "/proc"
locking_type = 1
wait_for_locks = 1
fallback_to_clustered_locking = 1
fallback_to_local_locking = 1
locking_dir = "/run/lock/lvm"
prioritise_write_locks = 1
abort_on_internal_errors = 0
detect_internal_vg_cache_corruption = 0
metadata_read_only = 0
mirror_segtype_default = "raid1"
raid10_segtype_default = "raid10"
use_lvmetad = 0
}

activation {
checks = 0
udev_sync = 1
udev_rules = 1
verify_udev_operations = 0
retry_deactivation = 1
missing_stripe_filler = "error"
use_linear_target = 1
reserved_stack = 64
reserved_memory = 8192
process_priority = -18
raid_region_size = 512
readahead = "auto"
raid_fault_policy = "warn"
mirror_log_fault_policy = "allocate"
mirror_image_fault_policy = "remove"
snapshot_autoextend_threshold = 100
snapshot_autoextend_percent = 20
thin_pool_autoextend_threshold = 100
thin_pool_autoextend_percent = 20
use_mlockall = 0
monitoring = 1
polling_interval = 15
}

metadata {
}

dmeventd {
mirror_library = "libdevmapper-event-lvm2mirror.so" snapshot_library = "libdevmapper-event-lvm2snapshot.so" thin_library = "libdevmapper-event-lvm2thin.so"
}