« first day (4885 days earlier)      last day (26 days later) » 

12:14 AM
A chairde - Morning all!
Wordle 1,066 5/6

⬛⬛⬛🟨🟨
🟨⬛⬛🟩⬛
⬛🟨⬛⬛⬛
⬛🟩🟨🟩🟩
🟩🟩🟩🟩🟩
 
Wordle 1,066 5/6*

⬜⬜⬜🟨⬜
🟨⬜⬜⬜🟨
⬜🟩🟩🟩⬜
⬜🟩🟩🟩🟩
🟩🟩🟩🟩🟩
 
 
2 hours later…
1:51 AM
@PaulWhite thx, it was BRINE
@PaulWhite do you save the screenshots or is there some way to got to a previous day?
 
2:25 AM
@HannahVernon I save them
Not for their own sake, rather as a side effect of sharing results and grids with others on a chat platform. I just need to scroll back a little way or search for the Wordle number.
 
 
9 hours later…
11:01 AM
Wordle 1,066 4/6*

⬛⬛⬛⬛🟨
⬛⬛🟨⬛⬛
⬛⬛🟨🟩🟩
🟩🟩🟩🟩🟩
 
 
3 hours later…
1:39 PM
well well well
still no frogs
should have stayed in bed
 
good night
 
1:50 PM
don't i wish
i have a call in 10 minutes
how will i cope
 
Is that a haiku?
 
close enough
an american haiku
 
You are a man of culture
 
> according to UniSuper's incident log, downtime started May 2, and a full restoration of services didn't happen until May 15.
oohwee
think they'll get a refund?
 
@PaulWhite I figured that you saved them to share ~18 hours later; with most of us being 12-18 hours behind NZ
 
i think that story is going to live forever as a reference for PMs pushing multi-cloud
> The joint statement and the outage updates are still not a technical post-mortem of what happened, and it's unclear if we'll get one.
i won't hold my breath
 
@HannahVernon The variously froggy text squares, yes. An image of my letters on the grid, no.
 
2:47 PM
Right
 
3:02 PM
@PeterVandivier egress fees are going to be a problem for many budgets, not to mention paying for the 2nd provider.
 
generally speaking, i find that that once the inevitability of cloud billing is accepted by an org, the run rate only ever goes up
and for those exceptions i know of that do keep cloud billing on a leash (at least that i've talked to so far), they've accept the additional cost as appropriate for the service rendered
 
I suppose one good thing is knowing that if you close your Google Cloud account, the data really is gone
 
😂💯
 
3:46 PM
i've got an interesting scenario where i think partitioning is fighting against optimize for sequential key
non-partitioned tables it works well for
this one partitioned table just keeps being sucky
 
repro or it didn't happen
 
4:29 PM
@ErikDarling Is it even functional with partitioned tables? I didn't think so, but come to think of it, I'm not really sure
The leading key column (in whatever your clustered index is) is the partition id, after all
 
@PaulWhite I’m not sure. I’m going to mess with it today and see if I can repro.
 
The absence of flow control waits would be a sign
 
I can see the btree flow control waits along with the latches and writelog
But the latch waits still dominate things
And the flow control waits are fairly high as well
 
So, it is functional at least
Just not effective
 
That seems to be the case
 
4:46 PM
What's the general scenario? Clustered index on an identity column? What sort of thing is the partitioning column? A datetime?
 
Correct
 
5:40 PM
I didn't know Boeing made helicopters
 
5:55 PM
ha ha
 

« first day (4885 days earlier)      last day (26 days later) »