> Dear Sean, We’re mere months away from PASS Summit and we want to know if you plan to attend! Take this (incredibly) short survey to let us know if you will be joining us, and if not, why we won’t see you at this year’s PASS Summit.
I won't be attending because @JoeObbish never setup a GoFundMe and PASS thinks that actual DBA things aren't worth the presentation space because...Linux.
I am going to fill this out and be brutally honest... like the kind of honest that would get me banned from this network
This is an option for why you're not going... omfg... this country.
"What if... guys, guys listen.. what if we make a datatype and call it long long int! Then we also call it uint64 and then we fill it with as many numbers as possible!"
@jadarnel27 Documentation is about all I can do, but I'd rather just fix the actual issue than document around it.
I'd put a uservoice item up for the standardization of units or to have the unit values added to the XE metadata. I'll upvote it and bring it up when/if I have a chance.
@JoeObbish, with minimally logged operations, the space allocations (changes to IAM, PFS, and GAM structures) are rolled forward and then rolled back if not committed so the operation is all-or-none. Changes made to data pages by bulk insert don't need to be rolled forward since those were physically written during the operation (which differs from normally logged operations where data file IO is asynchronous via lazywriter and checkpoint). — Dan Guzman12 hours ago
relevant to something that we discussed in here earlier
"There are two requirements for a small-query semaphore: The memory grant requested should be less than 5 MB The query cost should be less than 3 cost units"
otherwise it's not small
what do you mean by "scheduling action"?
you get a pair of semaphores for every resource governor pool
so the semaphore that you go to depends entirely on your pool and a few properties of the compiled query plan