cloudstack-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nitin Mehta (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CLOUDSTACK-5091) Upgrade to 4.2 - used cpu is getting bumped up when the over provisioning factor > 1 - script fix
Date Thu, 07 Nov 2013 23:27:18 GMT

    [ https://issues.apache.org/jira/browse/CLOUDSTACK-5091?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13816782#comment-13816782
] 

Nitin Mehta edited comment on CLOUDSTACK-5091 at 11/7/13 11:26 PM:
-------------------------------------------------------------------

Issue – Due to changes in over provisioning factor in 4.2, a bug got introduced for customers
using cpu.overprovisioning.factor (i.e. factor > 1)

Effects of the issue – You will see the used capacity bumped up by cpu.overprovisioning.factor
for all the vms deployed pre 4.2 version.

Who all need the fix – 
If you have upgraded to 4.2 or planning to upgrade to 4.2. AND
If cpu.overprovisioning.factor in the global settings is not set 1, then you need to apply
this fix.

Note – If you are planning to upgrade to 4.2.1 then this fix is not required.
When to fix
  If you are on 4.2 already then asap.
  If planning to upgrade to 4.2 then do it right after successful upgrade.

How to fix - 
Choose one of the fixes from below. Manual Fix or Script Fix

Manual Fix
Get over provisioning factor from query below. If factor = 1, move to Step #4 else Step #2.
select value from `cloud`.`configuration` where name='cpu.overprovisioning.factor';
Get a list of vm ids which need this fix. Move to Step #2
select id from `cloud`.`vm_instance` where removed is null and id not in (select vm_id from
`cloud`.`user_vm_details` where name='cpuOvercommitRatio')
For each vm in the list insert the over provisioning factors.
INSERT IGNORE INTO cloud.user_vm_details (vm_id, name, value) VALUES (?, 'cpuOvercommitRatio',
over_provisioning_factor_from_Step_#1)
INSERT IGNORE INTO cloud.user_vm_details (vm_id, name, value) VALUES (?, 'memoryOvercommitRatio',
1)
Verify from the dashboard that the used capacity has decreased by the over provisioning factor.
Done.


Script Fix
./fix_overprovisioning.sh -h
Usage: [-d] [-u] [-p]
-d - cloud DB server ip address, defaulted to localhost if not specified 
-u - user name to access cloud DB, defaulted to cloud if not specified 
-p - cloud DB user password, defaulted to cloud if not specified
Example: fix_overprovisioning.sh -dlocalhost -ucloud -pcloud
Verification – Notice the used capacity decreased by the over provisioning factor in the
dashboard view of the UI.


was (Author: nitinme):
Issue – Due to changes in over provisioning factor in 4.2, a bug got introduced for customers
using cpu.overprovisioning.factor (i.e. factor > 1)
Effects of the issue – You will see the used capacity bumped up by cpu.overprovisioning.factor
for all the vms deployed pre 4.2 version.
Who all need the fix – 
If you have upgraded to 4.2 or planning to upgrade to 4.2. AND
If cpu.overprovisioning.factor in the global settings is not set 1, then you need to apply
this fix.
Note – If you are planning to upgrade to 4.2.1 then this fix is not required.
When to fix
If you are on 4.2 already then asap.
If planning to upgrade to 4.2 then do it right after successful upgrade.
How to fix - 
Choose one of the fixes from below. Manual Fix or Script Fix
Manual Fix
Get over provisioning factor from query below. If factor = 1, move to Step #4 else Step #2.
select value from `cloud`.`configuration` where name='cpu.overprovisioning.factor';
Get a list of vm ids which need this fix. Move to Step #2
select id from `cloud`.`vm_instance` where removed is null and id not in (select vm_id from
`cloud`.`user_vm_details` where name='cpuOvercommitRatio')
For each vm in the list insert the over provisioning factors.
INSERT IGNORE INTO cloud.user_vm_details (vm_id, name, value) VALUES (?, 'cpuOvercommitRatio',
over_provisioning_factor_from_Step_#1)
INSERT IGNORE INTO cloud.user_vm_details (vm_id, name, value) VALUES (?, 'memoryOvercommitRatio',
1)
Verify from the dashboard that the used capacity has decreased by the over provisioning factor.
Done.
Script Fix
./fix_overprovisioning.sh -h
Usage: [-d] [-u] [-p]
-d - cloud DB server ip address, defaulted to localhost if not specified 
-u - user name to access cloud DB, defaulted to cloud if not specified 
-p - cloud DB user password, defaulted to cloud if not specified
Example: fix_overprovisioning.sh -dlocalhost -ucloud -pcloud
Verification – Notice the used capacity decreased by the over provisioning factor in the
dashboard view of the UI.

> Upgrade to 4.2 - used cpu is getting bumped up when the over provisioning factor >
1 - script fix
> -------------------------------------------------------------------------------------------------
>
>                 Key: CLOUDSTACK-5091
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-5091
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the default.) 
>    Affects Versions: 4.2.0
>            Reporter: Nitin Mehta
>            Assignee: Nitin Mehta
>             Fix For: 4.2.0
>
>
> Upgrade to 4.2 - used cpu is getting bumped up when the over provisioning factor >
1
> Provide a script fix for this - Will solve the issue where used cpu is getting bumped
up when the over provisioning factor > 1. This would need to be documented and run manually
as an additional step by the admin.
> Also write KB article



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message