stormseq search fails with JSON error about stormseq

[konradjk]

Strange. I think this is an Amazon issue, since it works if you use the Quick Launch Wizard instead of the Classic Wizard. I will look into it some more, but it should work if you go that route (note that you have to Edit Details on the last screen and change the Security Group to your stormseq group that has HTTP and HTTPS enabled). I will update the tutorial to use this and the updated Amazon interface.

[rghammonds]

Sorry, I don't understand "change the Security Group to your stormseq group that has HTTP and HTTPS enabled".

My choices are "Create New Security Group", or "Select Existing Security Groups", under which my choices are "default" and "quicklaunch-1".

I don't know what you mean by "your stormseq group". Something to do with the S3 bucket? Is it related to the .pem file that was downloaded when I first ran the quick launch wizard?

[rghammonds]

Ah, I never got to that step in the original directions. Clear now.

[rghammonds]

I've started an instance of stormseq, but it can't find my data file in the S3 bucket. Here's the S3 bucket

What should I enter in the stormseq "Amazon S3 bucket" and "Name your Genome" fields?

[konradjk]

In this case, enter gggh in S3 bucket, and ghexome in name your genome. You could also click multiple individuals and then only choose the ghexome sample in the pop-up.

[rghammonds]

That got a run started - it continued part way through the "right side" of the progress phase, then seemed to stop making progress. After a manual refresh the graphics disappeared. The new stats files in my S3 bucket seem to be empty. What happened?

[konradjk]

Did the entire "left side" (aligned, etc) get completely filled in or were there gaps? Was there any *.merged.bam file in the S3 bucket? Also, what version did you end up using?

[rghammonds]

The entire left side was filled in and the right side was partially filled in. There were no *.merged.bam files in the bucket. I used the latest version, stormseq_0.9 (ami-37e3935e).

[konradjk]

Were you using a spot instance? Most often if the process gets part of the way through the right side, but with no feedback, and then it fails, it is because the spot instances are terminated due to price. Unfortunately, we have no notification system in place yet to determine this. Immediately afterwards (I realize it's too late now, sorry about that), you can check the EC2 console to see if they were terminated by price, to confirm if that were the problem.

[rghammonds]

I wasn't using the spot instance.

On Tue, Jul 2, 2013 at 9:17 PM, Konrad Karczewski
[email protected]:

Were you using a spot instance? Most often if the process gets part of the
way through the right side, but with no feedback, and then it fails, it is
because the spot instances are terminated due to price. Unfortunately, we
have no notification system in place yet to determine this. Immediately
afterwards (I realize it's too late now, sorry about that), you can check
the EC2 console to see if they were terminated by price, to confirm if that
were the problem.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-20394773
.