- RSS Channel Showcase 3086235
- RSS Channel Showcase 5140928
- RSS Channel Showcase 4470557
- RSS Channel Showcase 6821206
Articles on this Page
- 06/17/16--19:24: _Comment on Read cur...
- 06/17/16--19:25: _Comment on Gluent N...
- 06/17/16--19:28: _Comment on Gluent N...
- 06/17/16--19:37: _Comment on Advanced...
- 06/21/16--12:28: _Comment on Gluent N...
- 06/21/16--13:18: _Comment on Gluent N...
- 06/21/16--13:28: _Comment on Gluent N...
- 06/22/16--16:06: _Comment on Gluent N...
- 07/02/16--07:51: _Comment on Advanced...
- 07/04/16--11:02: _Comment on Advanced...
Yep but only during new query execution, not for already (long-)running queries... unless they've changed it recently.
You'll have a chance to ask this question from James during the session!
Well the enterprises are changing. Microsoft Windows runs a Linux subsystem now, MSSQL runs on Linux as other examples :) Storage will go towards bunch of dumb "disks", everything else done in software - and more open source software over time. Everything going towards software-defined, the era of monolithic hardware and software solutions is over. Not all companies know that yet though. And IBM still does plenty of mainframe business, so the old world is not going away, will keep ticking on in the background.
AWR data does not contain the required level of detail that latchprofx gives. Namely the link between latch holding session, SQL active during hold and the kernel function responsible for taking the latch. AWR latch misses sections provide some instance level detail, but it's instance level so not always useful for drilling down into single user/application. ASH links together the latch waiter session, SQL and latch name so it's a start (and the blocking session often shows the latch holder).
So better to start from ASH, not AWR systemwide reports. There's plenty of ASH usasge information out there, Oracle docs and for example this by Kyle Hailey: https://sites.google.com/site/embtdbo/wait-event-documentation/ash---active-session-history#TOC-WAITS
It wasn't that useful a session to be honest..from 'technical knowledge' point of view.
It was more like a semi-marketing presentation of the D5 from EMC.
Thanks for the feedback.
My main aim when planning this session was to make everyone aware that the era of monolithic Fibre Channel SAN arrays is over (regardless if they are filled with flash or disks). Data storage is moving closer to CPUs, eliminating multiple bottlenecks and hops (latency) in between. DSSD guys are the leaders in that space (direct PCI-attached *shared* storage), so it's good to see what they're doing.
Btw I have one more "general" GNW planned, but after that I plan to split GNW into two streams, one is for DBA types (like myself) who want tech details & internals of new technologies - and another one for talking about what's possible with new technologies from enterprise architecture point of view (also myself now). So, different streams for different audiences.
thanks Tanel! Still appreciate you efforts in getting this webinar setup.
I am looking forward to further GNW sessions and the wealth of knowledge coming from GNW.
Sure, no problem! We'll have more fun!
Fantastic! Tool came in very handy to me. Much appreciated. I will have learn VBscript now :-)
I hoped to port the VBScript to python for use on other platforms, but never got to it.
A more modern way for CPU analysis (with Perf on Linux) would be this: