Follow-up Comment #2, bug #93993 (project xrootd):
I'm not sure my comment made it so I'll post it again.
As it works out the readv count was never included in the read count. So, we
were under counting. That makes it interesting to correct. I would suggest an
alternative to your proposal that better fits the way the data is actually
presented. Ideally,
ops.rd would only contain kXR_read requests (that's what we report)
ops.rdv would only contain kXR_readv requests
ops.rds would contain the total number of readv segments
We can call it ops.rdu if has better appeal; I'm neutral.
Given that we never added readv to read to backward compatible fix is to do
so and if you want to know actual reads you would need to subtract readv's.
That's cumbersome in a way. But given that we never reported them there is a
case for not adding them.
In any case. the above proposed way of doing it allows you to calculate just
about anything. So, please sign off or complain as I am in a position to add
this as part of the "f" stream monitoring changes.
_______________________________________________________
Reply to this item at:
<http://savannah.cern.ch/bugs/?93993>
_______________________________________________
Message sent via/by LCG Savannah
http://savannah.cern.ch/
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1
|