Version 0.1.2 released!

[Christopher Groskopf]

Users,

PANDA Project version 0.1.2 is now released. This version corrects for several oversights in the previous 0.1.1 release. All users are strongly encouraged to upgrade. (If you haven't already applied the 0.1.1 upgrade, do that first, then continue with the 0.1.2 upgrade.)

Upgrade documentation is available here:

http://readthedocs.org/docs/panda/en/0.1.2/upgrades.html

For a complete list of changes see the CHANGELOG:

http://readthedocs.org/docs/panda/en/0.1.2/index.html#id2

Please let us know how your upgrade goes and send us any other feedback.

Best,
Chris

[Sharon Machlis]

The upgrade went fine thanks, despite the numerous warning messages I
saw flash by the terminal (it might be helpful to mention in the
future if those are expected so we don't worry when we see them?).

Is 1.2 more resource-intensive than 1.0? I'm running a test
implementation using an AWS micro instance. I know that's not
recommended for production use, but for now we've just got a dozen or
so small data sets and 3 active users (although I set up some other
accounts that haven't been used yet). I don't want to commit to a $900/
year unless I'm sure folks here want and would use this, which is why
it's still on a micro instance. It seems significantly slower now than
1.0 was.

Not sure if anyone else is testing on a micro instance and noticed the
same slowdown?

On Apr 9, 5:19 pm, Christopher Groskopf <staringmon...@gmail.com>
wrote:
<snip>

[Christopher Groskopf]

Sharon,

> The upgrade went fine thanks, despite the numerous warning messages I
> saw flash by the terminal (it might be helpful to mention in the
> future if those are expected so we don't worry when we see them?).

What warning messages? Without seeing them I don't know if they were
expected or not...

Can you send me the contents of /var/log/panda-upgrade.log?

> Is 1.2 more resource-intensive than 1.0? I'm running a test
> implementation using an AWS micro instance. I know that's not
> recommended for production use, but for now we've just got a dozen or
> so small data sets and 3 active users (although I set up some other
> accounts that haven't been used yet). I don't want to commit to a $900/
> year unless I'm sure folks here want and would use this, which is why
> it's still on a micro instance. It seems significantly slower now than
> 1.0 was.

In a manner of speaking, yes. We doubled the amount of RAM dedicated
to the search engine, which on a micro is basically all the RAM on the
system. I suspect that as a result your system is swapping to disk
much more frequently, which would make everything slow. We do provide
some documentation that loosely explains how to scale that up and
down:

http://panda.readthedocs.org/en/latest/performance.html

However, I would just reiterate that we have no plans to support micro
instances, so the problem is likely to only get worse with future
releases. For example, we also introduced import throttling and
various other features into this release which may be aggravating the
problem.

Best,
Chris

[Sharon Machlis]

Actually, I'm now having a problem with the upgrade, I'm unable to add
data to an existing dataset. I just keep getting the error message

Sorry, this request could not be processed. Please try again later.

I don't know if it's because I'm running a micro instance or something
else. It's a small spreadsheet with 15 records to add to a set with
less than 1,000 records.

[Christopher Groskopf]

Hmmm, that sounds like an issue with the upgrade and not the instance
size. I notice in you original message you said you went from "0.1.0"
to "0.1.2". Did you do the 0.1.1 upgrade in between? They need to be
run in sequence.

Chris

[Gordon]

Chris,
   We just installed v 1.2 here at GGMC and are unable to add users.

   The page /admin/auth/user/add/ returns a HTTP 500 error. Other admin add pages (Categories, Tasks, Api keys) seem to work OK

Gordon

[Joe Germuska]

On Apr 20, 2012, at 1:46 PM, Gordon wrote:

   We just installed v 1.2 here at GGMC and are unable to add users. 

   The page /admin/auth/user/add/ returns a HTTP 500 error. Other admin add pages (Categories, Tasks, Api keys) seem to work OK

Alas, this is a known bug that has been fixed in the repository, but not in any release.

If this is seriously cramping your style, it's possible to work around it—I did, with some guidance from Chris. It's not novice level stuff, but I really needed to make it work because I had users I needed to add.

Joe

-- 

Joe Germuska

J...@Germuska.com * http://blog.germuska.com * http://twitter.com/JoeGermuska    

"I felt so good I told the leader how to follow."

-- Sly Stone 

[Andy Boyle]

We've got 1.0 installed and whenever we add a user it just takes me to a blank page. It adds the user, but the page that pops up is just blank. So it's easy enough for me to fix.

Is this a known issue?

-Andy

[Sharon Machlis]

Meanwhile, is there a way to revert back to 1.0? I'm finding 1.2 on a
micro instance to be pretty much unusable, even with a small amount of
data. I can't ask our research manager to test it like this, and I'm
not ready to ask for a budget commitment for a larger instance until
people have tried it out and know they find it useful. If and when we
wanted to use it for our newsroom, I'd upgrade to the current version
and a small instance.


On Apr 20, 1:35 pm, Christopher Groskopf <staringmon...@gmail.com>
wrote:

[Christopher Groskopf]

Sharon,

Sorry for the delayed reply.

There is currently no way to revert back to an early version, or I should say there is no prescribed way. It's technical possible, but there is no automation and I wouldn't recommend it. If it's crucial that you continue testing on a micro and don't mind giving me HTTP & SSH access to your PANDA I can try to tune it down so it will be functional on that hardware, but it would be a short-term fix at best.

Chris

[Christopher Groskopf]

Andy,

I don't believe that is a known bug for 1.0. There is an issue with adding users in 1.2, but I don't remember there being an issue in 1.0. If it is successfully adding the user I'd suggest (and my peril) that you just keeping doing what you're doing until you can safely migrate to a newer version. If you've got less-proficient users who need to use it and it's interfering then let me know and I'll be happy to debug it in place.

Best,

Chris

[Sharon Machlis]

As of today, the file I tried to add to my existing data sets in my
post-upgrade Panda appears in the list of data files. However, I also
see a message

Queued for import!
Data is waiting to be made searchable. Search results may be
incomplete.

and in fact search results do not include what's in that additional
file. That file only had around 20 records in it and I uploaded it
days ago. I uploaded a similar file to that same data set a few weeks
ago in version 0.1.0 and it worked fine.

[Sharon Machlis]

Happy to say that Chris tracked down an upgrade bug that had nothing
to do with running on a micro instance, and our PANDA is back and
better than ever :) Thanks, Chris!

Sharon

On Apr 23, 9:34 am, Christopher Groskopf <staringmon...@gmail.com>
wrote: