These are chat archives for ManageIQ/manageiq/performance

8th
Sep 2015
Alex Krzos
@akrzos
Sep 08 2015 14:56
@dmetzger57 FYI for RHEV-large initial inventory 11minutes spent gathering the inventory, looking at the rhevm httpd log can clearly see it is hitting individual api for each object (vm, host, etc)
Dennis Metzger
@dmetzger57
Sep 08 2015 15:08
@akrzos how large is the RHEV large inventory
Alex Krzos
@akrzos
Sep 08 2015 15:08
3k vms
Jason Frey
@Fryguy
Sep 08 2015 15:10
Yeah, they don't have batch API for detailed information
Unless they've changed that recently
We use a pool of 8 threads and try to parallelize as much as possible
@dmetzger57 Another place skeletal refresh might be a win :)
Alex Krzos
@akrzos
Sep 08 2015 15:11
explains the multiple threads I saw under the /proc/pid/status
Dennis Metzger
@dmetzger57
Sep 08 2015 15:15
@Fryguy agreed, the big change will be a big win
Jason Frey
@Fryguy
Sep 08 2015 15:16
Here's where the inventory collection happens
(Also, I know it doesn't belong there...I keep forgetting to move it)
Alex Krzos
@akrzos
Sep 08 2015 15:54
@Fryguy Has anyone ever opened a RFE BZ against RHEV for a batched detailed information?
Jason Frey
@Fryguy
Sep 08 2015 16:39
don't know