Done.
I ran up on to this issue when was trying to deploy CoreOS with rich
cloud-init data. So yeah, I think in future will be more users who will
use "big" userdata.
Issue: https://issues.apache.org/jira/browse/CLOUDSTACK-8644
On 2015.07.16. 19:01, Pierre-Luc Dion wrote:
Martins,
post will work, if there is no feature request on cloudmonkey could you
add one in Jira ? I found too that it could be a valuable feature to
cloudmonkey.
ref:
http://docs.cloudstack.apache.org/projects/cloudstack-administration/en/4.5/virtual_machines.html#user-data-and-meta-data
Cheers,
On Thu, Jul 16, 2015 at 9:31 AM, Mārtiņš Jakubovičs <mart...@vertigs.lv>
wrote:
Looks like API calls with GET method can have only 4KB of userdata, but
with POST it can be till 32 KB.
https://github.com/apache/cloudstack/blob/master/server/src/com/cloud/vm/UserVmManagerImpl.java#L3434
CloudMonkey use GET method, when change method in source code to POST it
is possible to send API calls with bigger userdata.
On 2015.07.16. 15:31, Mārtiņš Jakubovičs wrote:
Hello,
I try to push 8KB user-data when deploy new instance though cloudmonkey,
but I receive error message:
Error 431: User data is too long for an http GET request
cserrorcode = 4350
errorcode = 431
errortext = User data is too long for an http GET request
uuidList:
Looks like CloudStack can't handle so big request, but as documentation
describe, it should handle till 32KB.
http://cloudstack-administration.readthedocs.org/en/4.3/api.html
ACS 4.3.2.
Best regards,
Martins
