The Official Status Thread
-
@M_Adams said in The Official Status Thread:
tomorrow’s 10h E2E/UAT (end 2 end / user acceptance testing) meeting.
A ten hour meeting? Tell me they're at least providing lunch and dinner.
-
@HardwareGeek said in The Official Status Thread:
Truth in TV. That's a pretty accurate description of the world.
-
@Benjamin-Hall said in The Official Status Thread:
Got caught in a <abbr title=torrential downpour, rather"">rainstorm
I see that you also got caught by the "clicking abbr with a single word selected puts the cursor before the quotes instead of inside".
I mean, it's not like this bug was reported (at least once, by me) more than a year ago. Bikeshedding emojis is more important, I guess?
-
Status: made some refactoring, ran the code again, got different results from the reference case that I ran just before starting. Oops. Reverted my changes, ran the code again, still got different results.
-
Status: Warp Charge 30 lives up to its reputation. Goddamn, that's fast.
-
STATUS
Someone has just learned about lamdas and now everything is a 20 line lambda "one liner".
-
@Applied-Mediocrity said in The Official Status Thread:
I mean, to start page numbering from anything different than 1, there's an 11 step tutorial.
- Click on the "Insert" tab
- Open the "Page Numbers" dropdown button
- Click on "Format Page Numbers"
- Click the "Start at..." radio button
- Type your desired starting page number
- Or you can use the spinner to set the number
- Accidentally click "Cancel" which loses your changes
- Good thing you have 11 steps
- Repeat steps 1-5
- Instead of step 5 you can still do step 6
- Hit OK to save your changes.
-
@hungrier said in The Official Status Thread:
@Applied-Mediocrity said in The Official Status Thread:
I mean, to start page numbering from anything different than 1, there's an 11 step tutorial.
- Click on the "Insert" tab
- Open the "Page Numbers" dropdown button
- Click on "Format Page Numbers"
- Click the "Start at..." radio button
- Type your desired starting page number
- Or you can use the spinner to set the number
- Accidentally click "Cancel" which loses your changes
- Good thing you have 11 steps
- Repeat steps 1-5
- Instead of step 5 you can still do step 6
- Hit OK to save your changes.
- ...
- Profit!
-
@remi said in The Official Status Thread:
Status: made some refactoring, ran the code again, got different results from the reference case that I ran just before starting. Oops. Reverted my changes, ran the code again, still got different results.
And after spending almost all day on that, I've unilaterally decided that my code is good and that I must have fucked up while running the reference case.
'cause, let's face it, what's most likely? That I fucked up reorganizing a couple of hundreds LoC across several functions, or clicking on a single button in the UI?
Also, given my state today it looks like whatever @HardwareGeek caught is contagious through TD... I hope it won't get as bad as it did for him...
-
@M_Adams The ship has some sort of wacky "gradual dissipation into space" mechanism you can employ during deceleration to avoid that. Which you can choose not to employ as a blow-up-a-planet weapon. Or, apparently, reconfigure in a few minutes to cut off at 97% complete and somehow focus the remaining 3% perfectly to employ as a tac nuke.
-
@hungrier said in The Official Status Thread:
- Instead of step 5 you can still do step 6
- Hit OK to save your changes.
- Mother... shitter... son of an... ass! *bam* *bam* Libre ate all header and footer formatting, and screwed up TOC offsets and alignment, because it's not compatible with DOCX format I have to send for review.
-
Status: Our SDN firewall badly needed new hardware, so I bought it a new server and scheduled replacement for tomorrow evening.
I need to install one update so I can get the version up to exactly the ISO download version from the website before saving a final backup to be restored on the new hardware.
Now's a good time to do that, so I let it update and:Uhoh. Reboot and the thing is stone dead. BIOS for the raid card just PANICs when it tried to boot. Re-seat every thing and it's still dead.
New server installed and slightly-the-wrong-version config loaded back in just fine. Everything seems to work!
I just can't believe that timing for a complete hardware failure, one day before it was going to be replaced anyway.
Edit: Good reminder on backups I guess, RAID doesn't save you when the controller card explodes. I wonder if the disks are corrupt or not, it would be interesting to know but I don't have another one of those System x RAID cards.
-
Have you learnt nothing from the discussions here?
"If it ain't broken, don't update it."
-
@Zerosquare
"And even if it is, why take the chance?"
-
@Cursorkeys said in The Official Status Thread:
RAID doesn't save you when the controller card explodes.
That's why I prefer Software RAID
-
@TimeBandit said in The Official Status Thread:
That's why I prefer Software RAID
Hardware RAID is almost certainly just software RAID on a dedicated embedded processor. Which might be fine, or might be just waiting to go boom…
-
@Zerosquare said in The Official Status Thread:
"If it ain't broken, don't update it."
@izzion said in The Official Status Thread:
"And even if it is, why take the chance?"
Ahem:
"If it aint' broke, fix it until it is"
Also, the website said 'thou shalt not restore a backup made on an earlier version/patch level'
@TimeBandit said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
RAID doesn't save you when the controller card explodes.
That's why I prefer Software RAID
The lack of battery backup for the cache worries me on software RAID, power failures do happen often enough for that to potentially be a problem.
-
In my experience, RAID has brought down more servers than it's ever saved. I prefer simple disk setups with robust nightly backups. If a disk fails, swap it out and restore a backup image, in about 5% of the time it takes to try to get a failed RAID array going again.
-
@Cursorkeys said in The Official Status Thread:
The lack of battery backup for the cache worries me on software RAID, power failures do happen often enough for that to potentially be a problem.
In my experience with ext4 on RAID-1, you don't have any issue with power failures. When you reboot, it just reads the journal and apply what's missing. It only takes a couple seconds. YMMV
-
@TimeBandit said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
The lack of battery backup for the cache worries me on software RAID, power failures do happen often enough for that to potentially be a problem.
In my experience with ext4 on RAID-1, you don't have any issue with power failures. When you reboot, it just reads the journal and apply what's missing. It only takes a couple seconds. YMMV
But the write cache will be in system RAM, so that just gets lost. Or am I misunderstanding how it works?
-
@mott555 said in The Official Status Thread:
I prefer simple disk setups with robust nightly backups.
RAID isn't a substitute for backups. Totally different use case. RAID lets you build very large virtual disks (and possibly handle hardware failures without having to start from scratch with a rebuild). Backups let you recover from catastrophe, including user-driven catastrophes (such as plain old deleting a critical file).
-
@dkf
What do you mean, it's right there in the name!
Redundant
Array of
Interchangeable
DisksSo, since they're redundant and interchangeable, you should never have to worry about failures!!!11eleven
-
@Cursorkeys said in The Official Status Thread:
But the write cache will be in system RAM, so that just gets lost. Or am I misunderstanding how it works?
Anything not yet written to storage that will survive a power cut isn't actually saved. It can be in the journal log, but just being in RAM isn't enough (unless that's backed by enough independent, working battery power to be effectively in NVRAM). This is why any database transaction with writes in it is typically expensive to commit: the DB is typically configured to not complete the commit until it is sure that the data won't go poof due to an ill-timed power glitch…
-
@dkf said in The Official Status Thread:
@mott555 said in The Official Status Thread:
I prefer simple disk setups with robust nightly backups.
RAID isn't a substitute for backups. Totally different use case. RAID lets you build very large virtual disks (and possibly handle hardware failures without having to start from scratch with a rebuild). Backups let you recover from catastrophe, including user-driven catastrophes (such as plain old deleting a critical file).
In theory. In practice, this has never worked for me. Any kind of failure, even in a supposedly fault-tolerant RAID setup such as RAID-1 or RAID-5, has resulted in me having to restore everything from backup to a new array/disk after spending fruitless hours trying to get the RAID array to "just rebuild" after swapping out the one failed disk. It's so much easier to skip the RAID and go straight to backup instead. I've also had RAID arrays fail for no discernible reason, afterwards the controller checks out and all disks check out but the array is toast and needs to be remade from scratch.
We don't require super-high uptime, though. If things go down for a couple of hours, it doesn't grind our business to a halt.
(Never meant to call RAID a replacement for backup. What I meant is I've had to restore from backups because of RAID far more often than on non-RAID systems.)
-
@mott555 As far as I'm aware, the RAID arrays I've built haven't failed yet anyway. It helps a lot that they're using good parts, mostly low-traffic, and kept in relatively benign environments (no power, cooling or major dust problems).
-
@dkf said in The Official Status Thread:
no power
Hardware doesn't fail (detectably) if it has no power.
no power ... problems
Oh. That works too, maybe, I guess.
-
Status: EF Code First Migrations take twice as long to set up sane default column widths than a standard Database project & schema compare deploy. And counting...
-
@HardwareGeek said in The Official Status Thread:
Oh. That works too, maybe, I guess.
It's still emailing me every month to say “still working, no failures yet” so I'm guessing no problems…
-
@dkf said in The Official Status Thread:
@topspin said in The Official Status Thread:
and then a year or two later a (maybe) searchable index site will be added that links you to specific time codes in the video
Those links will work for a few hours, and then MS will reorganise the way all the links on their site work, again, and the timecode links will cease to be relevant. They'll probably reorder the frames in the video too.
No, the links will remain exactly the same. But the video will be replaced and reorganized.
-
@HardwareGeek said in The Official Status Thread:
Hardware doesn't fail (detectably) if it has no power.
Oh yeah?
-
@mott555 said in The Official Status Thread:
In theory. In practice, this has never worked for me
One day I should retell how I spent two weeks literally hot-re-plugging disks to reset their controllers because the server was hammering them too hard, and how such an action didn't result in data loss...
-
Status: I like living on the bleeding edge.
In the current FreeNAS nightly, apparently, starting jails will lock up the UI such that many functions (i.e. listing the jails, listing storage pool details, etc) just spin.
I believe this is because one of the jails fails the batch operation but doesn't complete the operation.
Oh, and generating a debug dump isn't working at present because of unrelated issues, and that's naturally the first thing the devs ask for...
-
Status: I think I'll have to give up on PyCharm.
I just want to edit a few simple Python scripts, but it does nothing but index shit all the time. Right now I can't even open a file I want to edit because it's busy "loading" the directory tree. There's only a few files and subdirectories directly in there. Those subdirectories themselves are large, but I didn't ask it to look in there. So for the last 10 minutes it did this:Compare with:
$ time ls study_ap4 <snip 5 lines> real 0m0.013s user 0m0.000s sys 0m0.002s
Compared to just using a plain text editor, the time I saved with its nice IDE features is more than made up by the time I lost dealing with this bullshit.
-
@topspin Side Bar WTF > New Topic > "PyCharm 2019 I fucking hate you" thread is ...
-
@topspin I sense that I'm not the only one who thinks jetbrains products are overrated.
STATUS I'm getting too old for midweek drinking.
@Applied-Mediocrity said in The Official Status Thread:
@topspin Side Bar WTF > New Topic > "PyCharm 2019 I fucking hate you" thread is ...
Change it to jetbrains and I'll chip in a rant.
-
@DogsB said in The Official Status Thread:
I sense that I'm not the only one who thinks jetbrains products are overrated.
I've always loathed them, but that's because they're somehow incompatible with how I work. Never analysed what that boils down to exactly; I just hate using them.
-
Wanted to buy State Fair tickets. Advance purchase discount was expiring last night. When I went online to try to buy them, the ticket purchasing site was completely unresponsive. According to comments on the Fair's Facebook, it had been down for much of the day. The Fair says they'll extend the advance purchase deadline; however, as of this morning, the site's back up, but it doesn't let you select tickets. The Fair's first day starts in about fifteen minutes. This is the first year that the Fair has used whatever ticketing platform fell the fuck over.
-
-
@Cursorkeys said in The Official Status Thread:
I wonder if the disks are corrupt or not
In any case, don't leave them plugged into the controller or they definitely will be
-
Status: TFW an elevator that said it's going down, went up instead.
TFW when after it reached floor, it didn't even open doors and immediately went down.
-
@ben_lubar said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
I wonder if the disks are corrupt or not
In any case, don't leave them plugged into the controller or they definitely will be
Whole thing has just gone in the skip after hitting the disks with a hammer, definitely corrupt now!
IBM does make lovely machines, everything in it was in little sections with easy removal handles, and status LEDs everywhere (lightpath diagnostics). A work of art almost.
-
I can't believe it's been almost two years since I was last on here. Having a good employer does weird things to you ...
-
- took my mom shopping
- forgot my trunk was full of weird toys and bondage gear
-
@abarker said in The Official Status Thread:
I can't believe it's been almost two years since I was last on here. Having a good employer does weird things to you ...
Welcome back!
-
@error Don’t tell her about “lifestyle”.
-
Luckily I had just enough time to conceal it all.
-
@Gąska said in The Official Status Thread:
Status: TFW an elevator that said it's going down, went up instead.
TFW when after it reached floor, it didn't even open doors and immediately went down.
Sorry, sorry! Been working on the code still. Making an elevator do the right thing is actually kinda complicated!
-
@error Concealed Carry?
I'm pretty sure brandishing them would be effective in stopping crime in progress.
-
@abarker said in The Official Status Thread:
I can't believe it's been almost two years since I was last on here. Having a good employer does weird things to you ...
Holy shit!
-
@Tsaukpaetra said in The Official Status Thread:
Making an elevator do the right thing is actually kinda complicated!
It actually isn't, but apparently that doesn't stop elevators being stupid.: