Creating Jobs is broken (NaN for timeouts)

Hi,

I just tried to create a job, but I can't get past the initial create job form (i.e., to the selection of examples / provers). Every value for wallclock/cpu timeout is rejected with "NaN second max timeout" (even if it's the default).

Marc

Sorry about that, Marc. We'll investigate and hopefully patch this today.
Aaron

mabrocks wrote:Hi,

I just tried to create a job, but I can't get past the initial create job form (i.e., to the selection of examples / provers). Every value for wallclock/cpu timeout is rejected with "NaN second max timeout" (even if it's the default).

Marc

Hello,

Sorry about the problem you're having creating jobs. This bug can occur if no queue is selected on the form. When you try to create a job, can you confirm that a queue is selected, and if so, could you say which one it is? Thanks for the help.

Hi Eric,

[quote="eric-burns"]

mabrocks wrote: This bug can occur if no queue is selected on the form. When you try to create a job, can you confirm that a queue is selected, and if so, could you say which one it is? Thanks for the help.

Ah, this pointer was helpful: No queue is selected, and the drop-down menu for choosing queues is empty, so I cannot select one.

Marc

mabrocks wrote:Hi Eric,

eric-burns wrote:

mabrocks wrote: This bug can occur if no queue is selected on the form. When you try to create a job, can you confirm that a queue is selected, and if so, could you say which one it is? Thanks for the help.

Ah, this pointer was helpful: No queue is selected, and the drop-down menu for choosing queues is empty, so I cannot select one.

Marc

All right, thank you. The reason this is happening is that we've implemented queue-specific timeout limits. If no queue is selected, the page can't find any timeout limit, and you get the NaN error.

We're looking into why no queues are available to select right now, and we'll let you know when it has been resolved.

Ok, we fixed the problem. It looks like we have a bug in our admin interface where if we remove global access from one queue (as I did with all2.q this morning), somehow all queues are affected. I have restored global access for all.q and long.q now, so they are appearing again in the list of queues on the job creation page.

Sorry for the problem,
Aaron