[linux-lvm] Some questions after starting to use 1.1-rc1
Heinz J . Mauelshagen
mauelshagen at sistina.com
Thu Apr 25 05:26:01 UTC 2002
On Sat, Apr 20, 2002 at 05:00:08PM +0100, Piete Brooks wrote:
> I've been using 0.9.1_beta6 for ages, and just upgraded to 1.1-rc1
> (I like "-qn" and the fact that minor device numbers are static!)
>
> 1) I've been comparing the speed of LVM over RAID0 with striped LVM, using:
> perl -e '$n=1900;$x="X"x(1024*1024);foreach $i(1..$n){print $x}'
> and I find that striped LVM takes 50% longer.
> Due to the different areas on the physical disk, there should be some
> difference (hdparm shows 37MB/s vs 31MB/s) and the stride might effect
> things, but vmstat shows that with striped LVM there is a period of disk
> IO, then no disk IO at the end (sample at end of email). Any ideas why ?
Hi Piete,
which stripe size did you use?
Were the PVs on non-saturated pathes?
Did you compare LVM1 Raid0 to MD Raid0 on the very same hardware?
>
> 2) When I tried to extend the striped partition, it failed when it used up the
> first two PVs, and will not start the next two. Ideas ?
That's a LVM1 constraint which will disapear with LVM2.
>
> # lvextend -L+1 /dev/vg97/test
> lvextend -- rounding relative size up to physical extent boundary
> lvextend -- rounding size to stripe boundary size
> lvextend -- extending logical volume "/dev/vg97/test" to 2 GB
> lvextend -- not enough free/allocatable physical extents to extend logical
> volume "/dev/vg97/test"
>
> # pvscan
> ...
> pvscan -- ACTIVE PV "/dev/.../part26" of VG "vg97" [1020 MB / 0 free]
> pvscan -- ACTIVE PV "/dev/.../part27" of VG "vg97" [1020 MB / 1020 MB free]
> pvscan -- ACTIVE PV "/dev/.../part26" of VG "vg97" [1020 MB / 0 free]
> pvscan -- ACTIVE PV "/dev/.../part27" of VG "vg97" [1020 MB / 1020 MB free]
> pvscan -- total: 24 [39.13 GB] / in use: 24 [39.13 GB] / in no VG: 0 [0]
>
> #
>
> 3) When trying out the new "-qn" (rah rah!) I wanted to tell the system
> forget about vg99; stop one PV; vgscan
> to see how it got on with a missing PV. The only way I could think of to
> say "forget vg99" was to vgexport it. Stopped one RAID array, and vgscan
> found vg99 (no need for "-qn" as I remember). Great ! restarted the array.
> However, I could not the "forget" about vg99 again to allow me to add the
> PV back again, as when I tried to export it, it said it was already
> exported (I assume because it saw the array which I had restarted ...
> a) Is there a better way to say "forget about vg99" ?
> (e.g. "vgremove --leave-all-data-ASIS /dev/vg99")
> b) should it be possible to export a VG if there is a PV of that VG which
> is still exported ?
No need to vgexport to test it.
Just stop one PV as you did adn "vgscan;vgchange -ay vg99" and it should
complain about gone VG quorum.
"vgchange -qn -ay vg99" should help this situation and activate the VG.
>
> 4) http://www.sistina.com/lvm_howtos/lvm_howto/Recovery_when_things_go_w.html
> (etc) are empty apart from:
> This is just stuff from the end of Joe's new LVM HowTo draft that I
> didn't know where to put. It should get moved to separate files as
> content is added
> but there is no "stuff" :-(( Where is "Joe's" HowTo ?
You found an old reference :-(
We'll remove it.
>
> 5) is there some utility to display the contents of VGDA on disk, /etc/lvm*,
> etc; show where the first PE is in a PV, etc ?
Run "pvdata -a" on the particular PV.
>
>
> vmstat output:
>
> procs memory swap io system cpu
> r b w swpd free buff cache si so bi bo in cs us sy id
> 1 0 0 0 1658156 6640 323648 0 0 0 0 135 53 0 50 50
> 1 0 1 0 1488868 6640 464556 0 0 0 60832 751 42 0 84 16
> 1 0 1 0 1394820 6640 553644 0 0 0 53640 872 37 0 82 18
> 1 0 1 0 1325356 6640 631468 0 0 0 43532 935 32 1 78 21
> 2 0 1 0 1269744 6640 689836 0 0 0 54848 1095 31 0 100 0
> 1 0 1 0 1192132 6640 759468 0 0 0 59680 925 32 0 100 0
> 2 0 1 0 1163724 6640 808620 0 0 0 44600 1139 31 0 100 0
> 1 0 1 0 1125648 6640 854700 0 0 0 55364 1625 35 1 99 0
> 1 0 1 0 1085020 6640 896684 0 0 0 60380 2188 31 0 100 0
> 1 0 1 0 956880 6640 975532 0 0 0 91424 1639 30 0 100 0
> 1 0 1 0 885072 6640 1061548 0 0 0 36624 916 34 1 77 22
> 1 0 1 0 870856 6640 1104556 0 0 0 38384 1169 31 0 99 0
> 1 0 1 0 762148 6640 1190692 0 0 0 69908 873 31 0 93 7
> 1 0 1 0 711120 6640 1259180 0 0 0 37896 1331 32 1 87 12
> 2 0 1 0 655180 6640 1315500 0 0 0 57808 1356 34 1 99 0
> 1 0 1 0 576164 6640 1381036 0 0 0 67068 1261 33 0 100 0
> 2 0 1 0 557596 6640 1423020 0 0 0 43252 1170 29 1 99 0
> 2 0 1 0 444896 6640 1500144 0 0 0 82376 1372 34 0 100 0
> 1 0 1 0 411028 6640 1549996 0 0 0 45384 1112 33 0 100 0
> 1 0 1 0 359984 6640 1604268 0 0 0 55176 1157 33 2 97 1
> 1 0 1 0 322548 6640 1654276 0 0 0 90664 1264 35 0 95 5
> 1 0 1 0 256308 6640 1725492 0 0 0 68960 1147 35 0 56 43
> 1 0 1 0 168056 6640 1813744 0 0 0 42748 941 139 0 49 51
> 1 0 1 0 108168 6640 1873632 0 0 0 76124 1120 126 0 51 49
> 0 1 1 0 35012 6640 1946788 0 0 0 45216 928 174 0 46 54
> 1 0 1 0 17240 6640 1964560 0 0 0 52916 1183 315 0 51 49
> 1 0 1 0 17244 6640 1964556 0 0 0 67456 1149 34 0 56 44
> 1 0 0 0 17244 6640 1964556 0 0 0 59724 838 36 0 54 46
> 1 0 0 0 17244 6640 1964556 0 0 0 0 767 31 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 112 31 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 117 31 0 51 49
> 1 0 0 0 17244 6640 1964556 0 0 0 0 115 30 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 109 33 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 116 31 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 109 30 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 110 33 0 51 49
> 1 0 0 0 17244 6640 1964556 0 0 0 0 111 31 0 50 50
> 1 0 0 0 17244 6640 1964556 0 0 0 0 115 31 0 50 50
> 1 0 0 0 17240 6640 1964560 0 0 0 0 114 32 0 50 50
> 1 0 0 0 17240 6640 1964560 0 0 0 0 116 31 0 51 49
> 1 0 0 0 17240 6640 1964560 0 0 0 0 114 32 0 52 48
> 1 0 0 0 17240 6640 1964560 0 0 0 0 116 31 0 50 50
> 1 0 1 0 50584 6640 1933328 0 0 0 0 124 43 0 53 46
>
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm at sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
--
Regards,
Heinz -- The LVM Guy --
*** Software bugs are stupid.
Nevertheless it needs not so stupid people to solve them ***
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Heinz Mauelshagen Sistina Software Inc.
Senior Consultant/Developer Am Sonnenhang 11
56242 Marienrachdorf
Germany
Mauelshagen at Sistina.com +49 2626 141200
FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
More information about the linux-lvm
mailing list