Failed to create volume Insufficient disk space.

[fed.linuxgossip]

Hello,

My applogic grid is showing : 59 GB of free space, when i try to create a new application with 40GB disk, it says Insufficient disk space.

Memory 7.89GB total (3.77GB free)
Storage 276.35GB total (59.98GB free)


As a matter of fact, I have only 2 vds nodes, the 1st one was 40GB, second one did not allow 40GB with the insufficient disk error and created it with 35GB space.


Example:-
------------------

grid1> app provision GSC_1_2_1_1 vdscustomer3 hostname=vds01.testing.com primary_ip=y.y.y.y netmask=255.255.255.240 gateway=x.x.x.x dns1=z.z.z.z dns2=z1.z1.z1.z1 user=vds2
user_pw=e$r5&8ddW root_pw=D#8%KmNG&&%^@FcX67 cpu.dflt=1 mem.dflt=1024M GSC.boot.size=40G .user1=testing .user2=vds3 .description="vds01.testing.com"
Provisioning application vdscustomer3...
Failed to create volume vdscustomer3.class.GSC.boot - Insufficient disk space.
Failed to create destination volume.

Failed to copy volume: GSC.boot

grid1>



Please advise.

[Karl]

The 59GB free, is there a 40GB chunk of that on one hardware node? As AFAIK you need that space free on a single node, not made up of little bits of storage on various nodes.

[muppie]

then the applogic software should really tell you the "smallest" (or biggest depending on the way you look at it) chunk in addition to the total aggregate free space

[fed.linuxgossip]

my provider is going to contact with 3tera on storage pool size, looks like there is some issue.

[Karl]

I actually forgot the most obvious thing, the free space shows just that - total free space, not taking account of mirroring. So if you want to have two copies of a volume then you need to halve the value, three copies then divide by 3 etc. The reason being, is that each volume can have a differing number of mirrors, so it can't really tell you anything other than total free space without use of any mirroring. Also, don't forget that the management node takes up space equivelant to the largest volume you want to create, for the impex volume. So if you wanted to create volumes with a max size of 80GB, and you had two appliances using 40GB volumes then with 2 way mirroring you'd use:

2 x 80GB = 160GB (For impex)
2 x (2 x 40GB) = 160GB (For appliance volumes)
= 320GB

Or that's my understanding of it anyway.

[fed.linuxgossip]

What do you make out of this ?

grid1> vol list --all
Scope Name Size(MB) Server State Filesystem Mirrored Mount State
-----------------------------------------------------------------------------------------------------------
gc_cpanel GSC.boot 40960 srv3,srv4 ok ext3 Y in_use
gc_cpanel1 GSC.boot 35840 srv2 degraded ext3 Y in_use
GSC_1_2_1_1 GSC.boot 2048 srv1,srv2 ok ext3 Y available
Lamp_1_0_3_1 fs 50 srv3,srv4 ok ext3 Y available
Lamp_1_0_3_1 mon 50 srv1,srv2 ok ext3 Y available
Lamp_1_0_3_1 mysql 100 srv3,srv4 ok ext3 Y available
/system HLB.boot 80 srv1,srv2 ok ext3 Y available
/system HLB.usr 200 srv3,srv4 ok ext3 Y available
/system IN.boot 80 srv1,srv2 ok ext3 Y available
/system IN.usr 200 srv3,srv4 ok ext3 Y available
/system INSSL.boot 80 srv1,srv2 ok ext3 Y available
/system INSSL.usr 200 srv3,srv4 ok ext3 Y available
/system LINUX.boot 2048 srv1,srv2 ok ext3 Y available
/system LINUX5.boot 2048 srv3,srv4 ok ext3 Y available
/system LUX.boot 80 srv1,srv2 ok ext3 Y available
/system LUX.usr 200 srv3,srv4 ok ext3 Y available
/system LUX5.boot 100 srv1,srv2 ok ext3 Y available
/system LUX5.usr 120 srv3,srv4 ok ext3 Y available
/system MON.boot 200 srv1,srv2 ok ext3 Y available
/system MON.usr 500 srv3,srv4 ok ext3 Y available
/system MYSQL.boot 120 srv1,srv2 ok ext3 Y available
/system MYSQL.usr 200 srv3,srv4 ok ext3 Y available
/system MYSQL5.boot 120 srv1,srv2 ok ext3 Y available
/system MYSQL5.usr 180 srv3,srv4 ok ext3 Y available
/system NAS.boot 90 srv1,srv2 ok ext3 Y available
/system NAS.usr 200 srv3,srv4 ok ext3 Y available
/system NET.boot 80 srv1,srv2 ok ext3 Y available
/system NET.usr 200 srv3,srv4 ok ext3 Y available
/system OUT.boot 80 srv1,srv2 ok ext3 Y available
/system OUT.usr 200 srv3,srv4 ok ext3 Y available
/system PS8.boot 80 srv1,srv2 ok ext3 Y available
/system PS8.usr 200 srv3,srv4 ok ext3 Y available
/system RPL.boot 80 srv1,srv2 ok ext3 Y available
/system RPL.usr 200 srv3,srv4 ok ext3 Y available
/system WEB.boot 160 srv1,srv2 ok ext3 Y available
/system WEB.usr 400 srv3,srv4 ok ext3 Y available
/system WEB4.boot 160 srv1,srv2 ok ext3 Y available
/system WEB4.usr 400 srv3,srv4 ok ext3 Y available
/system WEB5.boot 160 srv1,srv2 ok ext3 Y available
/system WEB5.usr 400 srv3,srv4 ok ext3 Y available
_SYSTEM boot 1022 srv1,srv2 ok ext3 Y in_use
_SYSTEM impex 16383 srv1,srv2 ok ext3 Y in_use
_SYSTEM meta 1023 srv1,srv2 ok ext3 Y in_use

grid1>

[PeterNic]

muppie - thanks for the suggestions; I'll open a change request to be able to show the largest individual allocation that can be made (e.g., the largest volume, the largest CPU allocation, etc.); most of this information is available internally, so we can expose it.

Update: the change request is SCR2563, "-Product: report the largest possible resource allocations". When it is implemented, this reference number will be included in the detailed set of changes in the release notes.