jasig-ssp IRC Logs-2013-08-07

Owned by Apereo Infrastructure
Last updated: Aug 07, 2013
5 min read

[11:26:53 CDT(-0500)] <pspaude1> Ok, Dan I created the pull request. I'm a bit rusty, but it appears to me that it worked correctly. I pointed it to the ssp-master branch.

[11:27:18 CDT(-0500)] <dmccallum54> thx. i'll check

[11:31:35 CDT(-0500)] <dmccallum54> 31 unread messages in inbox…. really looking forward to that

[11:31:59 CDT(-0500)] <dmccallum54> my guess is the vast majority are unsolicited notifications re how well SSP is working

[11:33:13 CDT(-0500)] <pspaude1> Yes it should be.

[11:36:44 CDT(-0500)] <TonyUnicon> sorry my internet was out, had no way to dial into the standup

[11:37:29 CDT(-0500)] <TonyUnicon> I'm finishing up SSP-845, then I'll get back around to SSP-521

[11:37:44 CDT(-0500)] <TonyUnicon> did we want to talk about SSP-521 at some point today?

[11:40:18 CDT(-0500)] <JasonElwood> yes, based on Dan's availability. I'm on a call until 1p then from 2:30-3:30

[11:40:59 CDT(-0500)] <TonyUnicon> sorry about earlier, was not expecting an outage

[11:47:24 CDT(-0500)] <dmccallum541> want to chat when that call wraps up around 1p, then? i assume that's EDT? (10a MST?)

[11:49:29 CDT(-0500)] <TonyUnicon> 11

[11:49:32 CDT(-0500)] <JasonElwood> yeah

[11:49:40 CDT(-0500)] <TonyUnicon> 10 minutes in otherwords

[11:49:42 CDT(-0500)] <TonyUnicon> ?

[11:49:46 CDT(-0500)] <TonyUnicon> the works

[11:50:07 CDT(-0500)] <TonyUnicon> that works*

[11:51:09 CDT(-0500)] <JasonElwood> 54545 bridge?

[11:51:16 CDT(-0500)] <dmccallum541> sure

[12:26:30 CDT(-0500)] <dmccallum541> possibly related to SSP-521, or at least another dead-ending scenario to check: https://issues.jasig.org/browse/SSP-530

[12:28:27 CDT(-0500)] <dmccallum541> here's the ticket for the side-effects of a lookup by school_id resulting in excessive warnings about overwriting an existing record: https://issues.jasig.org/browse/SSP-515

[12:30:51 CDT(-0500)] <dmccallum541> the ticket says the way to fix it would be to introduce a lookup API without side-effects, which I still think is probably a good idea in general. but i think it's probably also true that the UI needs to be a bit smarter. I about knowing when it should throw itself into edit mode… specifically if an internal person record is found when you press 'retrieve' from external. it's been a long time since i've been in that code but

[12:34:00 CDT(-0500)] <TonyUnicon> ok

[18:17:44 CDT(-0500)] <pspaude1> Hey dan I do think we have a problem with search. I've had my suspicions the last couple days but I think I'm closer to what is happening (the actual issue not the cause).

[18:18:20 CDT(-0500)] <pspaude1> So I have the user I've been testing these reports with and a couple others. usually they disapper over days, today it happend within a few hours.

[18:18:49 CDT(-0500)] <pspaude1> I select * from db and he is there, but the search tool doesn't pick him up.

[18:19:40 CDT(-0500)] <pspaude1> The object status is 2?!?

[18:38:25 CDT(-0500)] <dmccallum54> sweet

[18:38:33 CDT(-0500)] <dmccallum54> nobody else is testing?

[18:38:55 CDT(-0500)] <pspaude1> testing what?

[18:39:18 CDT(-0500)] <dmccallum54> sorry… nobody else is testing in the same env and deleting your users out from under you?

[18:39:29 CDT(-0500)] <pspaude1> Nope its my local unless you hacked in (smile)

[18:40:12 CDT(-0500)] <dmccallum54> shhhh

[18:40:19 CDT(-0500)] <pspaude1> The docs person has spend a bit of time the last couple of days on Linux CI but she couldn't pinpoint anything and Jason has altered it a bit, so I thought all was well, until I couldn't find my users.

[18:40:35 CDT(-0500)] <dmccallum54> um

[18:40:46 CDT(-0500)] <dmccallum54> do you have query logging enabled in your local postgres instance?

[18:41:08 CDT(-0500)] <pspaude1> So, I'm tyring to think how its getting set. No, that is probably the next step.

[18:41:08 CDT(-0500)] <dmccallum54> if not, definitely do that so when it happens again we can go back and try to get some idea of the context in which the deletion occurred

[18:41:43 CDT(-0500)] <pspaude1> Definately

[18:42:07 CDT(-0500)] <dmccallum54> where else have we done work recently that would PUT to a person record...

[18:42:41 CDT(-0500)] <dmccallum54> or… where else in the app have you been working just prior to this happening… only in search?

[18:43:45 CDT(-0500)] <pspaude1> I haven't done anything with search or PUT, but those are what I was thinking. Awhile ago I went through the latest commits to see if anything changed but I didn't see anything jump out.

[18:43:53 CDT(-0500)] <dmccallum54> there has been some recent work in the external person sync job, right? hard to see how it could cause people to be deleted, but if it is responsible it would definitely have this black-magical quality you describe

[18:44:24 CDT(-0500)] <dmccallum54> you've been testing reports? what have you been doing to test those?

[18:46:13 CDT(-0500)] <pspaude1> Not much, play around in ireport and then take the compiled files (jasper) and put them in the reports folder. Refresh SSP, search for my student and then go to action plan.

[18:47:43 CDT(-0500)] <pspaude1> Its a simple job, but iReport likes to blow up the whole report once you change enough, so its a back and forth game.

[18:48:21 CDT(-0500)] <dmccallum54> sounds fun

[18:48:34 CDT(-0500)] <dmccallum54> what're you doing in action plan? just adding some tasks, i assume?

[18:49:33 CDT(-0500)] <pspaude1> Yeah action plan was almost done and looked great, but then it got hung up on the condition for only displaying goals and strengths on the page. Really irks me because it was almost done.

[18:50:02 CDT(-0500)] <pspaude1> But, now as I go back to it, I added some more goals and tasks and the original report may not even be robust enough.

[18:52:28 CDT(-0500)] <pspaude1> The other one took longer as I had to add early alert created date. But, it is done and it had way more modifications. That was the coach history.

[18:55:30 CDT(-0500)] <dmccallum54> ok, well. i'm going to suggest turning on query logging and see if it happens again. hopefully it will. if it does happen again, but query logging is too much of a firehose to make heads or tails of the context, i'd probably suggest turning off the background jobs to try to narrow down the problem.

[18:56:09 CDT(-0500)] <pspaude1> Yeah that sounds like a good idea. I'm sure it will happen again.

[18:56:25 CDT(-0500)] <dmccallum54> that's good if you can at least get it happen frequently

[18:56:32 CDT(-0500)] <dmccallum54> flip those 2s back to 1s?

[18:57:00 CDT(-0500)] <pspaude1> Yes I did manually. There back and working.

[18:57:06 CDT(-0500)] <dmccallum54> cool

[18:57:41 CDT(-0500)] <pspaude1> I've also thought I saw status changes as well active to transition etc. So, I'll put logging on and keep my eyes peeled.

[18:59:36 CDT(-0500)] <dmccallum54> that 'strengths' tab in Action Plan is suspect b/c it's sending the entire person record, not just a strength

[18:59:47 CDT(-0500)] <dmccallum54> but so far i'm not able to reproduce what you're seeing

[19:00:40 CDT(-0500)] <dmccallum54> i'm sure you've been heads down, but any questions on the term start/end date stuff?

[19:03:06 CDT(-0500)] <pspaude1> Ok, I've had this happen in other tools like when I was doing early alert and journal tools. But, since it would be days between seeing it and trying to keep track of users between the two I would doubt what I was seeing

[19:03:24 CDT(-0500)] <dmccallum54> ugh

[19:03:26 CDT(-0500)] <pspaude1> not really

[19:03:40 CDT(-0500)] <pspaude1> Your comment is pretty comprhensive

[19:03:44 CDT(-0500)] <dmccallum54> heh. ok.

[19:05:07 CDT(-0500)] <pspaude1> I'll get back into it tmrw and see what happens. Tried starting it again today, but everything else going on derailed it.

[19:05:22 CDT(-0500)] <pspaude1> Found the mail exception issue

[19:06:40 CDT(-0500)] <pspaude1> Its the recipient email in message, needs to be valid.