Hide Forgot
Hi Bala, This bug is to track providing hardware diagnostics info on webUI. Lets discuss about it along with AB. Krishna
*** This bug has been marked as a duplicate of bug 485 ***
Harsha, This is not related to monitoring, this is related to performance diagnostics (network disks etc). This is more like qualifying the hardware that it is fit to be use. Or even when a customer comes back with performance issue we run a set of performance tests to narrow down the problem. Bug 485 is more like giving info on what hardware is being used and whether it is working or not etc.
diagnostics information is part of hardware monitoring even checking the hardware if its fit to be used, its not just clearly specified in that ticket. Single ticket is enough as both as they are related to hardware certification process where we recommend certain hardware and we have tools to pin point problems which are provided by the hardware vendors. *** This bug has been marked as a duplicate of bug 485 ***
> where we recommend certain hardware and we have tools to pin point problems > which are provided by the hardware vendors. Harsha, This is a feature through which we need following output. * dd output over glusterfs (with and without iflag/oflag=direct) * dd output over all the exports (with and without iflag/oflag=direct) These things are not given by h/w vendors. And this is *very* important for the support when there are performance issues (mainly over cifs/nfs). For example, we had a timeout errors through CIFS on a client (and server side /cifs/<volume> was all fine).. hence we doubted so many different things, had to change so many options in smb.conf etc, but finally realized that its one of the RAID controller giving issues (as in response time on that was low, and was giving 6MBps performance).. if we have this option, this was 10mins task, instead we had to spend more than 3-4hrs, and customer almost thought this stuff doesn't work.. We need diagnostic info, that too, very soon.
but how are you proposing to use it with backend? . Adding RAID controller based approach will give all this metrics, even there are options to check performance of the each disks. So i am not sure what you all are talking about here. Just a "dd" would give numbers from the client side. But taking numbers directly from the backend there are other various tools provided by hardware vendors itself to diagonize the problem which i will making the list. So i just wanted it to be tracked with a single bugid rather than have multiple places.
This feature is under discussion will be available post 3.1
Created attachment 418 [details] patch XTerm app-defaults which fixes the keybindings
Sorry, I hate when it does that. I commented on another bug and it changed bugs on me. This patch was for a different bug.
Created attachment 419
What the??? I know I was on the right bug this time. I wish I could just delete...