[
https://issues.apache.org/jira/browse/WHIRR-682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13497512#comment-13497512
]
Steve Loughran commented on WHIRR-682:
--------------------------------------
absolute limits for this user
{code}
$ nova absolute-limits
+--------------------------+-------+
| Name | Value |
+--------------------------+-------+
| maxImageMeta | 20 |
| maxPersonality | 5 |
| maxPersonalitySize | 10240 |
| maxSecurityGroupRules | 20 |
| maxSecurityGroups | 10 |
| maxServerMeta | 20 |
| maxTotalCores | -1 |
| maxTotalFloatingIps | 5 |
| maxTotalInstances | 100 |
| maxTotalKeypairs | 100 |
| maxTotalPrivateNetworks | 0 |
| maxTotalRAMSize | 66560 |
| maxTotalVolumeGigabytes | -1 |
| maxTotalVolumes | 0 |
| totalCoresUsed | 0 |
| totalInstancesUsed | 0 |
| totalKeyPairsUsed | 0 |
| totalPrivateNetworksUsed | 0 |
| totalRAMUsed | 0 |
| totalSecurityGroupsUsed | 0 |
| totalVolumeGigabytesUsed | 0 |
| totalVolumesUsed | 0 |
+--------------------------+-------+
{code}
none of these were exceeded.
The rate limits are a lot tighter
{code}
$ nova rate-limits
+--------+----------------+-------+--------+--------+--------------------------+
| Verb | URI | Value | Remain | Unit | Next_Available |
+--------+----------------+-------+--------+--------+--------------------------+
| DELETE | /servers | 4 | 4 | MINUTE | 2012-11-14T21:59:03.304Z |
| GET | * | 100 | 100 | MINUTE | 2012-11-14T21:59:03.304Z |
| GET | /servers | 25 | 25 | MINUTE | 2012-11-14T21:59:03.304Z |
| POST | * | 100 | 100 | MINUTE | 2012-11-14T21:59:03.304Z |
| POST | /os-networksv2 | 100 | 100 | DAY | 2012-11-14T21:59:03.304Z |
| POST | /servers | 4 | 4 | MINUTE | 2012-11-14T21:59:03.304Z |
| POST | /servers | 1000 | 1000 | DAY | 2012-11-14T21:59:03.304Z |
+--------+----------------+-------+--------+--------+--------------------------+
{code}
# There's a nominal limit on how many servers you can create of 4/minute,
though it seemed to kick in earlier.
# if there's really a limit on deletions of 4/minute, that's going to impact
deletion too.
There is a retry handler in the JCloud requester -you could have one that
parsed the exception, read its wait-time, slept until then and tried again
(with some jitter and backoff for parallel requests). But how to set that from
Whirr?
> OpenStack Nova throttling prevents Whirr from bringing up clusters with
> multiple nodes
> --------------------------------------------------------------------------------------
>
> Key: WHIRR-682
> URL: https://issues.apache.org/jira/browse/WHIRR-682
> Project: Whirr
> Issue Type: Bug
> Components: core
> Affects Versions: 0.9.0
> Environment: whirr.provider=rackspace-cloudservers-uk
> Reporter: Steve Loughran
>
> You can't create a 3 node hadoop cluster on OpenStack Nova, because it
> rate-limits VM requests (in time and per-day). The endpoints send a 413-coded
> message, but this is not detected as a retryable action
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
