Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

[15:19:40 CDT(-0500)] <sjungling> you recommend starting there and then going to more advanced pooling if the issue persists?

[15:20:48 CDT(-0500)] <EricDalquist> well that would fix the deferred issue

[15:20:51 CDT(-0500)] <EricDalquist> \but if queries are still sllow

[15:20:57 CDT(-0500)] <EricDalquist> you're just going to change the overload

[15:21:11 CDT(-0500)] <EricDalquist> to a potential large increase in the number of open connections

[15:21:23 CDT(-0500)] <EricDalquist> I'm still back on ... why are ldap queries taking so long

[15:21:36 CDT(-0500)] <EricDalquist> I'd assume that a direct user query should be like .... a millisecond or two

[15:21:46 CDT(-0500)] <sjungling> right

[15:21:47 CDT(-0500)] <EricDalquist> that is the whole point of LDAP versus a DB

[15:21:55 CDT(-0500)] <EricDalquist> if you're seeing this behavior with a 5s timeout

[15:22:10 CDT(-0500)] <EricDalquist> that seems to imply your LDAP server is having problems

[15:22:15 CDT(-0500)] <sjungling> right

[15:23:32 CDT(-0500)] <sjungling> 5s is the default timeout?

[15:23:41 CDT(-0500)] <sjungling> with the webproxy portlet, right?

[15:23:55 CDT(-0500)] <EricDalquist> I'm not sure, I'd have to go look at portlet manager

[15:31:36 CDT(-0500)] <EricDalquist> uhg athena I'm trying to track down this ALL stuff

[15:31:40 CDT(-0500)] <EricDalquist> such a rabbit hole

[15:32:30 CDT(-0500)] <EricDalquist> https://github.com/Jasig/uPortal/commit/b412d0f13fa336db165c2722c9fe05091d5602c7

[15:32:36 CDT(-0500)] <EricDalquist> that is the commit that did it though

[15:33:04 CDT(-0500)] <EricDalquist> https://github.com/Jasig/uPortal/commit/b412d0f13fa336db165c2722c9fe05091d5602c7#L6R88

[15:33:06 CDT(-0500)] <EricDalquist> that block

[15:33:23 CDT(-0500)] <EricDalquist> that ends up in a chunk of code that tries to figure out if the ALL_TARGET is a portlet or a person

[15:33:32 CDT(-0500)] <EricDalquist> and since it doesn't match the portlet pattern it assumes it is a PERSON

[15:33:36 CDT(-0500)] <EricDalquist> and runs it through PAGs

[15:33:43 CDT(-0500)] <athena> ugh

[15:33:43 CDT(-0500)] <EricDalquist> which does the attribute query

[15:34:31 CDT(-0500)] <EricDalquist> the thing is I can;t figure out where ALL_TARGET is actually ever tested against

[15:34:36 CDT(-0500)] <EricDalquist> but I haven't read the whole patch yet

[15:34:38 CDT(-0500)] <EricDalquist> have to run

[15:34:43 CDT(-0500)] <EricDalquist> and will dig into this more tomorrow.