[Planetlab-devel] Re: api sessions & other issues with the web UI
Faiyaz Ahmed
faiyaza at CS.Princeton.EDU
Fri Feb 23 12:12:04 EST 2007
Yes, a new rc candidate will likely be checked into CVS late today or
early next week after some preliminary testing. Clearly the website has
had a good number of fixes applied to it, the node software (especially
NM) has seen a great deal of additions, and a few other minor things
including but not limited to the Bootmanager and pl_mom have had some
work done to them.
It would only make sense to bump the release. I will email the list
when its ready.
F
Thierry Parmentelat wrote:
> Thanks Reid, it does help
> My repositories are uptodate - wrt HEAD. Should I use another update
> strategy to get these changes ?
> Btw, are there any plans to build an rc2 for this planetlab-4.0 ?
> Thierry
>
> Reid Moran wrote:
>> We already fixed this bug, update your repositories. The GetEvents
>> calls was somehow destroying each user's session when they accessed a
>> page making that call. I have changed the web pages so that only
>> admins can even access that call. This fixes the problem, we are
>> looking into the GetEvents call and redoing the API to make it work
>> better right now.
>>
>> Hope this helps.
>>
>> -Reid
>>
>> Thierry Parmentelat wrote:
>>> Hi Reid
>>>
>>> I'm experiencing a bug in the web UI. I'm having this issue on the
>>> one-lab site but not on planet-lab.org:
>>> Whenever a given node's details are displayed, the web session
>>> becomes useless; the api session gets rejected by the API and any
>>> given subsequent call to the API fails. It turns out the session
>>> entry in the database somehow gets wiped off.
>>>
>>> On my site I've found out that I could getting rid of the issue if I
>>> did not call GetEvents in nodes/index.php
>>>
>>> Does this ring any bell ? It sounds like an issue that appeared on
>>> your website a while ago, that you guys probably have fixed.
>>> ==
>>> On a similar track, I've found that UpdateNode.py needed a patch for
>>> regular users to be able to change their node's boot state. Again
>>> this is an issue that appeared in the support ML so I take it you've
>>> found out already.
>>>
>>> ==
>>> I've also tweaked the nodes browser to support patterns in hostnames.
>>> I am about to commit my changes, and will of course describe all
>>> these changes in the cvs log so you can keep up.
>>>
>>> Hope this helps - Thierry
>>
>
> _______________________________________________
> Devel mailing list
> Devel at lists.planet-lab.org
> https://lists.planet-lab.org/mailman/listinfo/devel
More information about the Devel
mailing list