Re: [ovirt-users] multipath configuration for local disks?

[Matthias Leopold]

Am 2017-10-10 um 12:53 schrieb Yaniv Kaul:



On Oct 10, 2017 12:23 PM, "Matthias Leopold" 
> wrote:


hi,

i'm using three different generations of hardware for my oVirt
hypervisor hosts. they are not exactly the "same", but very "similar".

- all were installed with oVirt Node 4.1.x installers
- all have (2-4) local SSD disks for oVirt Node OS
- all were configured with manual partitioning and SW RAID 1 with
two SSD disks for oVirt Node OS

after initializing the third generation of hosts (with 4 SSD disks)
i noticed the difference in multipath configuration for the local
disks. 



I may be wrong but it rarely matters on local disks. Unless they have 
some kind of active-active (multi-channel?) controllers.

Y.


thanks, you're probably right. i was asking
(a) for the sake of consistency
(b) out of curiosity
(c) because the multipathing adds an additional level of abstraction in 
the whole setup (which includes SW RAID)


strictly speaking none of these are show-stoppers , i'll manage...

matthias
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] multipath configuration for local disks?

[Yaniv Kaul]

On Oct 10, 2017 12:23 PM, "Matthias Leopold" <
matthias.leop...@meduniwien.ac.at> wrote:

hi,

i'm using three different generations of hardware for my oVirt hypervisor
hosts. they are not exactly the "same", but very "similar".

- all were installed with oVirt Node 4.1.x installers
- all have (2-4) local SSD disks for oVirt Node OS
- all were configured with manual partitioning and SW RAID 1 with two SSD
disks for oVirt Node OS

after initializing the third generation of hosts (with 4 SSD disks) i
noticed the difference in multipath configuration for the local disks.


I may be wrong but it rarely matters on local disks. Unless they have some
kind of active-active (multi-channel?) controllers.
Y.

suddenly the two OS disks (for RAID 1) are not configured as multipath
disks anymore, the unused other two are multipath disks. in the first two
host generations the (only) two OS disks (for RAID 1) are configured as
multipath disks.

after looking at this for the first time and briefly looking into multipath
documentation i think that local disks shouldn't be included in multipath
configuration at all. i tried to remove the local disks from multipathing
(delete from wwids file, explicit blacklisting), to no avail.

i'm pasting two listings from the 4 SSD disk situation:

# lsscsi --scsi_id -g
[0:0:0:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sda   -  /dev/sg0
[0:0:1:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdb   -  /dev/sg1
[0:0:2:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdc
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216  /dev/sg2
[0:0:3:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdd
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214  /dev/sg3

# ls -l /dev/disk/by-id/*SAMSUNG*
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J102212 -> ../../sdb
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J102212-part1 -> ../../sdb1
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J102212-part2 -> ../../sdb2
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J102214 -> ../../sdd
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J102216 -> ../../sdc
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J10 -> ../../sda
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J10-part1 -> ../../sda1
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/ata-SAMSUNG_MZ
7LM480HMHQ-5_S2UJNX0J10-part2 -> ../../sda2
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/dm-name-SAMSUN
G_MZ7LM480HMHQ-5_S2UJNX0J102214 -> ../../dm-6
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/dm-name-SAMSUN
G_MZ7LM480HMHQ-5_S2UJNX0J102216 -> ../../dm-5
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/dm-uuid-mpath-
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214 -> ../../dm-6
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 /dev/disk/by-id/dm-uuid-mpath-
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216 -> ../../dm-5

can someone explain this behaviour? what is the _intended_ configuration?

thx
matthias



___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] multipath configuration for local disks?

[Matthias Leopold]

hi,

i'm using three different generations of hardware for my oVirt 
hypervisor hosts. they are not exactly the "same", but very "similar".


- all were installed with oVirt Node 4.1.x installers
- all have (2-4) local SSD disks for oVirt Node OS
- all were configured with manual partitioning and SW RAID 1 with two 
SSD disks for oVirt Node OS


after initializing the third generation of hosts (with 4 SSD disks) i 
noticed the difference in multipath configuration for the local disks. 
suddenly the two OS disks (for RAID 1) are not configured as multipath 
disks anymore, the unused other two are multipath disks. in the first 
two host generations the (only) two OS disks (for RAID 1) are configured 
as multipath disks.


after looking at this for the first time and briefly looking into 
multipath documentation i think that local disks shouldn't be included 
in multipath configuration at all. i tried to remove the local disks 
from multipathing (delete from wwids file, explicit blacklisting), to no 
avail.


i'm pasting two listings from the 4 SSD disk situation:

# lsscsi --scsi_id -g
[0:0:0:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sda   -  /dev/sg0
[0:0:1:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdb   -  /dev/sg1
[0:0:2:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdc 
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216  /dev/sg2
[0:0:3:0]diskATA  SAMSUNG MZ7LM480 204Q  /dev/sdd 
SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214  /dev/sg3


# ls -l /dev/disk/by-id/*SAMSUNG*
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102212 -> ../../sdb
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102212-part1 -> 
../../sdb1
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102212-part2 -> 
../../sdb2
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214 -> ../../sdd
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216 -> ../../sdc
lrwxrwxrwx. 1 root root  9  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J10 -> ../../sda
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J10-part1 -> 
../../sda1
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/ata-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J10-part2 -> 
../../sda2
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/dm-name-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214 -> 
../../dm-6
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/dm-name-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216 -> 
../../dm-5
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/dm-uuid-mpath-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102214 
-> ../../dm-6
lrwxrwxrwx. 1 root root 10  9. Okt 12:07 
/dev/disk/by-id/dm-uuid-mpath-SAMSUNG_MZ7LM480HMHQ-5_S2UJNX0J102216 
-> ../../dm-5


can someone explain this behaviour? what is the _intended_ configuration?

thx
matthias



___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users