/dev/chat

3:12 AM

@derobert So copying and pasting isn't an option?

3:32 AM

The upshot of the message is that the DVD doesn't have a release file. It isn't really worth pasting. And I don't recall seeing such a message earlier.

What's the point of this? That using a DVD is inherently insecure?

Faheem Mitha

4:35 AM

Of course, it's probably been a while since I did a from-scratch installation from a CD/DVD.

derobert

4:53 AM

@FaheemMitha Not sure what message that is....

Faheem Mitha

@derobert I could post a screenshot.

Faheem Mitha

7:57 AM

I've got to say, gtklp's handling of the situation when you restrict to pages that are outside the range of pages of the document is really quite staggeringly unhelpful. Though I admit I'm not sure what optimal behavior would be. What does it actually do? It prints a sheet with the following on top:

<</.HWMargins[17. 12. 17. 12.]>>setpagedevice

I mean, seriously?

Actually, I suppose a good thing would be for gtklp to throw an error when the Print button is pressed.

Faheem Mitha

9:53 AM

Just heard back from someone who is probably the gtklp maintainer. He wrote:

> GtkLP is only a frontend for Cups and all print processing is done there.
So this is a problem with Cups, not GtkLP.

I suppose CUPS would have to throw an error for gtklp to throw an error.

Faheem Mitha

10:32 AM

Folks, does this look like a Nouveau driver problem?

terdon

@FaheemMitha No, it looks like an LVM issue, that's where the message appears, right?

It would help if we could see the final few lines though.

Faheem Mitha

@terdon I think it finished with LVM. It said:

> Activating lvm and md swap... done.

@terdon chat.stackexchange.com/rooms/62196/…

If you want to talk to the poster directly.

Personally, nouveau has my vote.

terdon

@FaheemMitha ah, yes, but that's a different image. The first had died at lvm

Faheem Mitha

@terdon Hmm. It's not clear, I'd say.

terdon

@FaheemMitha No, it isn't. But why are there two different images there? Both seem to be complaining about NULL Pointers but in one it happens when setting up LVM and in the other when trying to run fsck. Either way, those would seem to suggest an issue with the disks. What makes you think nouveau?

Faheem Mitha

10:47 AM

@terdon The call trace mentions nouveau.

And drivers, particularly reverse engineered ones, are relatively likely to cause a kernel oops.

terdon

@FaheemMitha Ah, yes it does.

Faheem Mitha

@terdon And I think the two images are two different attempts to boot. But you could ask the poster yourself. He's in that chat room right now.

terdon

And I've also seen this with graphics issues more than with anything else. But it does seem strange that it seems to have occurred when dealing with the hard drives in both cases.

Faheem Mitha

@terdon Well, it could just be that the next thing after the hard drive was the video card. But...<shrug>.

My experience with kernel panics is extremely limited. The kernel maintainers would probably have a better idea.

I've suggested to the poster that he post a bug report. But he doesn't seem keen.

Henry WH Hack v2.0

11:28 AM

@terdon This guy has posted a duplicate answer.

0

A: Error Installing mySql on Ubuntu 16.04

This is the correct solution for you First, you have remove all packages of mysql-server: sudo rm -rf /var/lib/mysql Then install: sudo apt-get install lamp-server^ Or you can do: sudo apt-get install mysql-server

I know this is not the AUGCR but.

terdon

@HenryWHHackv2.0 ?

Henry WH Hack v2.0

@terdon The answer was copied from another answer.

Maybe I'm wrong.

terdon

@HenryWHHackv2.0 Ah so not the same, but copied from someone else? Please flag it.

Henry WH Hack v2.0

As what?

terdon

Never mind, I deleted. But next time, flag for moderator attention and give us the link to the source they copied from.

Henry WH Hack v2.0

11:32 AM

Ok

terdon

It's better to flag because that way there's a record and other mods can deal with it if I'm not available etc.

Henry WH Hack v2.0

Ok

[Case Closed by]: HenryWHHackv2.0

derobert

4:53 PM

@FaheemMitha there is an awful lot of nouveau stuff in that backtrace for it to be unrelated

And I'm always suspicious of drivers I can't spell.

Faheem Mitha

5:48 PM

@derobert Drivers you can't spell? Meaning Nouveau? And there I thought you had French ancestry!

It's a bit weird to be telling people to use a proprietary driver, but pragmatism makes strange bedfellows. <Sigh.>

derobert

@FaheemMitha Italian :-/

Faheem Mitha

@derobert Oh, pardon me.

derobert

@FaheemMitha I'd try a newer kernel first. Presuming the person hasn't already tried the newest. Then report a bug.

Faheem Mitha

@derobert Good point. We were chatting in a room earlier. I could add that comment. Or you could comment on the question yourself.

I suppose that the Nouveau drivers would get updated along with the kernel.

derobert

Also, turns out half the reason I couldn't spell nouveau is that Chromium insists that it's not spelled "nouveau". Would I like to try Noumea?

Faheem Mitha

5:52 PM

Yes, I should I have thought to tell him to try a newer kernel. Still. a newer kernel wouldn't be supported either.

@derobert I wouldn't.

derobert

Faheem Mitha

That's the problem with a non-standard or custom kernel. You don't get bugfixes, security or otherwise.

derobert

... Chromium's idea of a video driver.

Faheem Mitha

@derobert Eww.

Hmm, a sea slug. A wonder of nature.

Looks like a surgery in progress.

derobert

Yep. Apparently video drivers come in yellow too:

Faheem Mitha

5:55 PM

@derobert Wow. Colorful.

derobert

@FaheemMitha Try the new kernel to see if it's fixed, if it isn't, report the bug upstream. If it is, report the bug to the distro—they need to pull the fix.

Faheem Mitha

@derobert Sounds good. But you're telling the wrong person. You're aware it's not my issue, right?

Do you want the link?

derobert

Yes. But I don't have the link :-P

Faheem Mitha

@derobert Do you want the link, or shall I just post what you wrote? Giving you proper attribution, of course. :-)

Parenthetically, apparently one place retired Indian film stars go where they retire is Parliament. Explains a lot, that.

derobert

@FaheemMitha Either is fine with me. Though it's certainly easier to let you post it.

Faheem Mitha

6:00 PM

2

A: Running a current Linux OS on older kernel (what are the / are there disadvantages?)

There's nothing wrong with using an older kernel, as long as the line is supported/maintained. In some case they are, most of the time they are not. And you do not want to use an older/unmaintained kernel, because it will not receive bug fixes, including security fixes. In the particular case of...

The poster is very polite, which makes a change from the usual.

BTW, I suggested he post the screenshot there. It's godawful, really. But blame me, not him.

And here is the room we were chatting in - chat.stackexchange.com/transcript/62196

In case you feel like adding something.

derobert

@FaheemMitha I'll let you post it...

Faheem Mitha

@derobert Um, ok. You mean a comment on the question?

derobert

@FaheemMitha yeah, on the question, or the chat room, wherever...

Faheem Mitha

Ok, will do.

derobert

@FaheemMitha also, another suggestion, since this is a server: possibly just use the svga or vesa driver instead. If it's indeed nouveau crashing, and it's not needed....

4.11.6-1~bpo9+1 is in stretch-backports

Faheem Mitha

6:08 PM

@derobert I suggested he try the proprietary drivers.

That seems like a more functional alternative, seeing as vesa is useless except for debugging. I mean, you don't want to run it.

derobert

That's only if you need the features (e.g., 3D accel, GPU compute). If the machine doesn't need that....

If it's a server, it may not even be running X.

Faheem Mitha

@derobert I think he just wants his machine to not crash. He does imply it's a server, it's true.

There's no real harm in using the Nvidia drivers, anyway. Unless you object to them on the grounds of software freedom.

Dumb question, but is the video card is responsible for everything displayed on the screen?

Even the scrolling text at the beginning?

(So, I'm relaying this comment from @derobert.) Try the new kernel to see if it's fixed, if it isn't, report the bug upstream. If it is, report the bug to the distro—they need to pull the fix. (credit to @derobert, naturally). Also, as I mentioned, try the proprietary Nvidia drivers. The problem is likely related to the open source Nouveau drivers, which are mentioned repeatedly in the call trace. — Faheem Mitha 6 mins ago

@derobert Hope that meets with your approval.

derobert

@FaheemMitha yes. You can continue to use the drivers its using at the beginning by not loading nouveau (or proprietary nvidia)

... and if you're not using X, that's likely fine.

Faheem Mitha

@derobert Ok, so some sort of rudimentary driver - vesa, perhaps.

But why is it crashing at the beginning, when X hasn't started up, though?

derobert

yeah, every PC video card has a minimal standard interface...

@FaheemMitha when you load nouveau, the kernel tries to switch to the nouveau frame buffer — away from the VGA one

that's always why, e.g., you'll suddenly see the text size change (typically allowing a lot more to show on screen)

Faheem Mitha

6:20 PM

@derobert Oh, and this can happen quite early in the boot?

@derobert Yes, I've noticed that. Interesting.

My loading nouveau, you mean the kernel module, presumably.

derobert

yes, the kernel module

Faheem Mitha

ok

derobert

a reasonable thing to try is just to blacklist it in modprobe and rebuild initramfs. Presuming X isn't used.

Faheem Mitha

@derobert You really should be making these comments. Is building initramfs necessary?

derobert

@FaheemMitha Not entirely sure if nouveau can be loaded by the initramfs or only later.

Faheem Mitha

6:31 PM

@derobert Oh.

Faheem Mitha

6:45 PM

@derobert So, thoughts about the annoying CUPS possible-bug I ran into? Mentioned in this room earlier. And I posted a question about it too.

derobert

@FaheemMitha that your printer spewed postscript as text?

Faheem Mitha

@derobert Or PDF. Yes.

It's nice to get some feedback on these things before posting a bug report. Which is likely to be ignored, anyway.

Assuming it's a bug, and not user error, of course.

I can see the devs saying - don't do that, then. Or words to that effect.

Random off-topic. This seems like horribly low pay. Unless South America is even cheaper than India.

38

Q: Client wants to reduce hourly rate at the start of a new project

I am a freelance developer with 3 years of experience in iOS and for receiving me as a systems engineer in December. I was working a month with a new client, we had set a rate of USD 20 per hour and 40 hours per month. There is no contract. After the month I sent the bill with 43 hours (I worked...

derobert

@FaheemMitha I haven't been paying much attention to the room... been busy today....

it was when you put in a silly page range?

Faheem Mitha

@derobert Ok. Well, I posted a question.

@derobert An incorrect range, yes.

derobert

7:01 PM

@FaheemMitha That strikes me as a bug. The desired behavior for invalid input is an error message, not weird output.

Faheem Mitha

@derobert Ok, thank you. Can you reproduce?

It's possible it's driver-specific or something.

Seeing as if it was a general problem, it would be likely to be fixed.

derobert

@FaheemMitha you're going to have to link me to the steps... also, I probably need to try at home, since here at work we have an actual PostScript printer.

Faheem Mitha

@derobert I just selected pages that don't exist in the document in gtklp. Unfortunately, gtklp saves those page ranges without giving any indication of the fact, which is part of the problem. Do you still use gtklp?

derobert

yeah... I can try at work.

ok, I just asked it to print pages 24–25 of a 3 page document

it made a job...

The job has failed

D [14/Jul/2017:15:09:18 -0400] [Job 565] Start rendering...
D [14/Jul/2017:15:09:18 -0400] [Job 565] Set job-printer-state-message to "Start rendering...", current level=INFO
D [14/Jul/2017:15:09:18 -0400] [Job 565] Processing page 1...
D [14/Jul/2017:15:09:18 -0400] [Job 565] Set job-printer-state-message to "Processing page 1...", current level=INFO
D [14/Jul/2017:15:09:18 -0400] [Job 565] No pages will be processed (FirstPage > LastPage).
D [14/Jul/2017:15:09:18 -0400] [Job 565] Rendering completed

Faheem Mitha

@derobert Yes, I got printer failed too. But did it print anything?

derobert

7:12 PM

@FaheemMitha No, don't think so. Didn't hear the printer warming up.

Faheem Mitha

Or did nothing happen?

@derobert Hmm, might be driver-specific then.

derobert

printer didn't warm up

Faheem Mitha

@derobert ok.

derobert

well, gtklp ought to have complained—it didn't. So there's an bug.

Faheem Mitha

@derobert The maintainer said it wasn't his problem. Or words to that effect.

derobert

7:13 PM

There might be a 2nd bug in your CUPS printer driver that caused it to actually print jumk

Faheem Mitha

I quoted him in the question.

@derobert Indeed. It's a Brother printer driver, unfortunately.

So not part of Debian.

derobert

Ah. I have a Brother printer at home. They are nice printers... if only they had equally-nice open source drivers.

Faheem Mitha

@derobert I couldn't agree more. :-)

On both counts.

It was also unfortunate timing. At the same time I left that setting in there inadvertantly, my printer decided there was no paper in the tray.

derobert

Found your question. I think the maintainer means gtklp essentially just builds a CUPS request without knowing at all what it means—so the error check ought to be in CUPS.

IOW, not an easy bug to fix :-(

Faheem Mitha

@derobert Well, the page range is set in gtklp.Though it's true that gtklp doesn't know the size of the document. But perhaps CUPS doesn't either.

derobert

7:25 PM

I don't think CUPS does at that point—not until it starts processing it. So probably not easy to fix.

Faheem Mitha

@derobert I suppose not. Should I delete that question? It doesn't seem to be getting much interest.

derobert

7:38 PM

I just voted to close it :-P

Faheem Mitha

7:57 PM

@derobert Et tu, Brute?

derobert