Skip to content

Conversation

@snuf
Copy link

@snuf snuf commented Jun 24, 2014

Pulling back up against master so we can keep following it a bit...

@yadvr
Copy link
Member

yadvr commented Dec 5, 2014

Hey, @snuf can you rebase this against master, let's merge it on master if it works functionally and the build passes with smoke tests?

@snuf
Copy link
Author

snuf commented Dec 6, 2014

Hey Rohit,

It’s been rebased with master two days ago, so should be a “breeze”, is there a specific point you want me to merge with ?

Cheers,

Funs

On 05 Dec 2014, at 18:31, Rohit Yadav [email protected] wrote:

Hey, @snuf can you rebase this against master, let's merge it on master if it works functionally and the build passes with smoke tests?


Reply to this email directly or view it on GitHub.

@yadvr
Copy link
Member

yadvr commented Dec 6, 2014

That's great Funs. I see it's already passing smoke tests on TravisCI so we can merge it on master whenever you advise? If it's functionally working with good enough coverage and tests I would say go ahead and merge it on master right away.

@snuf
Copy link
Author

snuf commented Dec 8, 2014

Hi Rohit,

Well there are a couple of things here:

1st: I work on a fork and pull in upstream master and merge that into ovmsupport (https://github.com/snuf/cloudstack/tree/ovmsupport) so in principle that makes the ovmsupport tree aligned with master including my patches.
2nd: Test coverage on part of the code is 81% (the object branch, supportive translation layer to communicate with Ovm3) and on part not where I want it to be (the hypervisor branch, where integration with Cloudstack lives).
3rd: It works within the boundaries of the feature design doc on wiki, which means all basic stuff works but some features can not be provided as Ovm3 doesn’t support them with the storage backend (running snapshots for example).
4th: The systemvm.img that’s generated from buildsystemvm.sh is a raw image so can be used for the systemvms for Ovm3.

Does this answer your question ? I genuinely think there should be some coverage for the integration layer into Cloudstack which is kind of absent?

Cheers,

Funs

On 06 Dec 2014, at 14:02, Rohit Yadav [email protected] wrote:

That's great Funs. I see it's already passing smoke tests on TravisCI so we can merge it on master whenever you advise? If it's functionally working with good enough coverage and tests I would say go ahead and merge it on master right away.


Reply to this email directly or view it on GitHub.

@yadvr
Copy link
Member

yadvr commented Dec 9, 2014

Hi Funs,

Thanks for replying. Alright let us know when you've good set of integration tests, I'm halfway with a new integration testing system semi-automated by ansible that uses CloudStack to test CloudStack. In that I run KVM, Xen etc. as VMs on KVM hosts all managed by CloudStack.

Regards.

Funs and others added 24 commits January 11, 2015 12:42
The User table's UUID column is restricted to 40 chars only, since we don't
know how long the nameID/userID of a SAML authenticated user will be - the fix
hashes that user ID and takes a substring of length 40 chars. For hashing,
SHA256 is used which returns a 64 char length string.

- Fix tests, add test cases
- Improve checkSAMLUser method
- Use SHA256 one way hashing to create unique UUID for SAML users

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit b2b4962)
Signed-off-by: Rohit Yadav <[email protected]>
As per Version 1 cookies, certain characters are now allowed such as space,
colons etc but they should be url encoded using UTF8 encoding. The frontend
has a cookie value unboxing method that removes any double quotes that are added.

As per the doc http://download.oracle.com/javase/6/docs/api/java/net/URLEncoder.html
values are application/x-www-form-urlencoded and as per
http://www.w3.org/TR/html4/interact/forms.html#h-17.13.4 whitespaces are encoded
as +, therefore '+' are replaced by %20 (whitespace).

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 734bd70)
Signed-off-by: Rohit Yadav <[email protected]>
The fix generates X509Certificate if missing from DB and uses that for eternity.
SAML SP metadata remains same since it's using the same X509 certificate and
it remains same after restarts. The certificate is serialized, base64 encoded
and stored in the keystore table under a specific name. For reading, it's
retrieved, base64 decoded and deserialized.

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 4358714)
Signed-off-by: Rohit Yadav <[email protected]>
…n.com

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 23de431)
Signed-off-by: Rohit Yadav <[email protected]>
yadvr and others added 9 commits February 27, 2015 18:24
- insecure authenticators excluded in configuration
- snapshot response should have zone
- remove vmsnapshots when removing accounts

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 5481485)
Signed-off-by: Rohit Yadav <[email protected]>

Conflicts:
	api/src/org/apache/cloudstack/api/response/VMSnapshotResponse.java
	server/src/com/cloud/api/ApiResponseHelper.java
	server/src/com/cloud/storage/download/DownloadActiveState.java
Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit dd5fb2e)
Signed-off-by: Rohit Yadav <[email protected]>
Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 814e557)
Signed-off-by: Rohit Yadav <[email protected]>
Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit f70afa1)
Signed-off-by: Rohit Yadav <[email protected]>
- Adds X-XSS-Protection header
- Adds X-Content-Type-Options header
- Fixes to use json content type defined from global settings
- Uses secure cookie if enabled in global settings

Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit b6b3494)
Signed-off-by: Rohit Yadav <[email protected]>
Signed-off-by: Rohit Yadav <[email protected]>
(cherry picked from commit 0f819f1)
Signed-off-by: Rohit Yadav <[email protected]>
git merge master
Merge remote-tracking branch 'cloudstack/master' into ovmsupport
@yadvr
Copy link
Member

yadvr commented Mar 2, 2015

@snuf Funs how about we merge your stuff in?

@snuf
Copy link
Author

snuf commented Mar 3, 2015

Hey Rohit,

Well all was fine until I did my last two merges from master after doing the 3.3.x integration of OVM, allocate vnet breaks and rolls back on the DB all of a sudden am trying to figure out where that is. So all bits work except for network provising for a VM (jeez luckily not important at all…), so vm rollout breaks. Am looking at where and why it breaks, but am pretty not impressed at the moment :\

Cheers,

Funs

On 02 Mar 2015, at 03:14, Rohit Yadav <[email protected] mailto:[email protected]> wrote:

@snuf https://github.com/snuf Funs how about we merge your stuff in?


Reply to this email directly or view it on GitHub #9 (comment).

@snuf
Copy link
Author

snuf commented Mar 3, 2015

Kay happy to report it was a problem in my rollout script which I changed recently cough so everything is fine, it’s just that the change left out the vlan range, so that explains a lot..

So you can go ahead and merge if you want to, I’m 3 days a way from master atm (need me to do a new pull request?).

Cheers,

Funs

On 02 Mar 2015, at 03:14, Rohit Yadav [email protected] wrote:

@snuf https://github.com/snuf Funs how about we merge your stuff in?


Reply to this email directly or view it on GitHub #9 (comment).

@snuf snuf closed this Mar 5, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.