dlcusa | rwp, I burned my 3.0.0 net-installer ISO onto the same USB-stick and still had the XZ issue on the first laptop (Dell Latitude E6410) but it booted as expected on the second (Acer Aspire One NAV50). So something has changed since 3.0.0, it appears. | 03:57 |
---|---|---|
gnarface | dlcusa: how did you burn the iso? dd? | 04:01 |
dlcusa | Yup | 04:01 |
gnarface | make sure for future burns you always run "sync" afterwards or add conv=fsync to the dd command-line | 04:02 |
gnarface | it could cause sporadic incomplete burns (unreported) that look like a corrupt iso | 04:02 |
gnarface | the acer doesn't look like it has enough ram for the live image, but the netinstaller should be fine | 04:03 |
gnarface | i'd try the memtest86 thing | 04:03 |
dlcusa | I'll let you know how that goes--going away for almost a week very soon. | 04:04 |
dlcusa | The hour of Dell diagnostics were perfect. | 04:05 |
gnarface | it's hard to say what could be the issue at this point it's a weird one | 04:06 |
gnarface | i'd even check to make sure the dell has the latest bios | 04:06 |
gnarface | of course you can still make an image that doesn't use xz to boot but for an i5 it really shouldn't have to come to that | 04:06 |
dlcusa | Will doublecheck that, but it's getting old. | 04:07 |
gnarface | yea but it's still newer than everything i have here | 04:07 |
gnarface | and they all still boot | 04:07 |
gnarface | and it shouldn't be new enough to need a bleeding-edge kernel | 04:07 |
dlcusa | Yeah, that's what I mean--the newer stuff is more suspect. | 04:08 |
gnarface | maybe the usb key is failing | 04:08 |
dlcusa | It's new. | 04:08 |
gnarface | hmmm | 04:08 |
gnarface | yea i got nothing. it's gonna require more deductive testing to diagnose | 04:09 |
gnarface | for the dell maybe the live image is worth a try since it has plenty of ram | 04:09 |
gnarface | i think you can find a live image with a newer kernel | 04:09 |
gnarface | just in case the 4.19 in beowulf has some rare error with that hardware | 04:10 |
gnarface | the acer is an Atom cpu so i have really no idea what to expect from it | 04:10 |
gnarface | word is they're supposed to work like regular intel chips but i've never actually seen it first hand | 04:10 |
dlcusa | Well, I expect this problem is going to eat up a lot of my time. :-( | 04:11 |
gnarface | i wish it wouldn't but you do know that if you can successfully use the older installer, it should be no different once upgraded after install, right? | 04:12 |
gnarface | actually debugging the newer installer may be an unnecessary distraction here | 04:13 |
dlcusa | Yes, but I really like to isolate issues, even if I'm the only person affected. | 04:14 |
gnarface | yea, i understand, and people here would like to know what's wrong too, but if you're pressed for time it's important to know that the result from upgrading to current from a 3.0.0 or 3.0.1 or 3.0.whatever iso should be identical after the fact | 04:15 |
gnarface | some of us have machines here that haven't seen an install disk since debian | 04:15 |
gnarface | heh, that didn't really sound as confidence-inspiring as i meant it to | 04:16 |
dlcusa | Well we're all resource-poor for the most part. | 04:16 |
gnarface | my point is just that the upgrade path should be complete regardless of which version you start from | 04:17 |
dlcusa | Yeah, I understand that. But the rescue disk needs to boot reliably. | 04:18 |
gnarface | well, yea i agree with that but you might have better luck generating one from a fully upgraded install | 04:18 |
dlcusa | Gotta run for a bit. | 04:19 |
rwp | I might be inclined to remove the USB, then attach it again, then read back the written data and verify that what was read was what should have been written. | 04:21 |
rwp | But yes buffered I/O not having gotten written yet might cause a problem. | 04:21 |
rwp | gnarface and I have disagreed on the best way to write USB devices in the past so I won't suggest anything but it can take several minutes to write and without O_DIRECT this might be in the background and not immediately indicative of when it is done writing. | 04:23 |
gnarface | hmm, it occurs to me to wonder how large the USB key is itself | 04:25 |
gnarface | since i've had problems with Dells having size limits | 04:25 |
gnarface | though also that should be for an earlier generation of hardware not relevant to this one | 04:26 |
gnarface | i'd still be wary if the usb key is very large, and i'd try again with a 2GB one if i had it on hand | 04:26 |
gnarface | (or failing that, an optical disk) | 04:27 |
gnarface | a dell here from the core2duo era can't boot anything physically larger than 2GB from SD or USB slot, even though after boot the hardware clearly can recognize full sdxc spec cards | 04:28 |
gnarface | the manufacturer's info doesn't bother to warn you that the bios can't before post | 04:28 |
gnarface | and that i5 is a much newer dell but i guarantee the bios source code came from the same origin | 04:29 |
gnarface | (and note: to be clear this is a fundamentally different issue from the previous generation "legacy bios 2gb boot partition" max limit issue, because partitioning a larger device down to 2GB doesn't help in this case) | 04:30 |
fluffywolf | I was helping someone on a welding forum with a similar issue... apparently if you use a SD card over 2GB for a firmware upgrade it bricks certain welders. lol | 04:30 |
gnarface | hah | 04:30 |
gnarface | ouch | 04:30 |
onefang | is that better or worse than welding a brick? | 04:31 |
gnarface | on that same dell i also had to resort to using lilo because for whatever reason grub takes half an hour to boot | 04:32 |
luser978 | Dells are always so "special". I'm sure they'll sell you a Dellium [tm] usb stick (capacity withheld, preinstalled image) which happens to work, for a firstborn or two. | 07:52 |
luser978 | Nobody buys Dell outside service contracts, it's Russian roulette frequently. | 07:53 |
luser978 | There are oceans of optiplex and so on low end refurbished workdtations on sale all the time dirt cheap, for a reason. | 07:55 |
luser978 | *sta* | 07:55 |
dlcusa | So I found a 2 GB USB-stick I could repurpose, put the 3.1.1 net-installer on it, and the Dell boots it. Thanks for the assistance, gnarface and rwp. | 08:06 |
golinux | Please take it to #devuan-offtopic | 08:17 |
rwp | dlcusa, Thanks for the closure on the problem. I have heard of 2GB limits before so this seems to be another one of those cases. I wouldn't have thought of it myself here though. gnarface got that one. Glad you got it going. | 08:29 |
dlcusa | Me, too, rwp--looks like I'll be taking the Dell with Beowulf installed for the trip after all! | 09:01 |
gnarface | dlcusa: wow shocking to hear their bios still has that limitation in such a relatively late model, but i'm glad we at least diagnosed the issue | 09:16 |
gnarface | dlcusa: if you want real irony, the optical drive will still probably boot fine from a 8GB dvd | 09:16 |
dlcusa | Well, progress is slowed when the PTB have agendas against some progess, right? | 10:02 |
dlcusa | Well, maybe not in this particular case, perhaps. | 10:03 |
ham5urg | Is there a GUI for adding a VPN in XFCE? | 15:35 |
sadoon_albader[m | You could use network manager's gui | 16:02 |
ham5urg_ | sadoon_albader[m, you mean to install package network-manager-gnome | 16:14 |
sadoon_albader[m | yup | 16:15 |
sadoon_albader[m | the nm-applet has that feature | 16:15 |
sadoon_albader[m | or nm-connection-editor | 16:16 |
ham5urg_ | nm-tray is a qt application. Is this the right choice for xfce? Couldn't find package nm-connection-editor | 16:19 |
gnarface | try "apt-cache search ^network-manager | 16:21 |
gnarface | " | 16:21 |
sadoon_albader[m | oh no | 16:24 |
sadoon_albader[m | nm-connection-editor is not a package | 16:24 |
sadoon_albader[m | it comes with network-manager-gnome | 16:24 |
Guest9 | Hi, anyone else having problems with beowulf 3.1.1 amd64 netinst accessing the mirror ? it seems like the "mirror" does not support the beowulf release... | 16:26 |
ham5urg_ | Does NM get listed in Xfce when adding it to the top-bar? Can't find it though | 16:26 |
onefang | Which "mirror"? We have a bunch of them. | 16:27 |
gnarface | Guest9: it was just working for someone else earlier | 16:27 |
onefang | deb.devuan.org is a round robin DNS that resolves to various of our mirrors. | 16:28 |
Guest9 | the mirrors to choose are "deb.devuan.org" and "pkgmaster.devuan.org", what is then resolved, idk. but i tcpdumped and saw that the installer stops after the webserver complains about the byterange requested. | 16:29 |
gnarface | weird | 16:29 |
gnarface | they both respond that way? | 16:30 |
gnarface | if you only tried deb.devuan.org once you might have just got bad luck; try again | 16:30 |
Guest9 | both show the same behaviour, but i only tcpdumped one of them. and i tried multiple times. | 16:31 |
onefang | Best to NOT use pkgmaster, coz the other mirrors sync to that, and things work better when everyone else isn't keeping it busy. | 16:31 |
onefang | https://pkgmaster.devuan.org/mirror_list.txt is a list of all our mirrors. | 16:31 |
Guest9 | onefang yes of course, but for testing why sth does not work ... | 16:32 |
onefang | Oh if you just are testing things, https://sledjhamr.org/apt-panopticon/results/Report-web.html might be useful. | 16:32 |
onefang | http://veritas.devuan.org/apt-panopticon/results/Report-web.html and https://borta.devuan.dev/apt-panopticon/results/Report-web.html run the same tests from other parts of the world. | 16:33 |
onefang | Probably in a couple of weeks time I'll be able to pull my finger out and fix up apt-panopticon. Currently the failures for deb.devuan.org are bogus, but I know why. | 16:36 |
onefang | But for now, I should sleep soon. lol | 16:40 |
Guest9 | thanks for the note about the mirror status page with apt-panopticon, now testing one that shows to be correct | 16:40 |
Guest9 | i tried to explicitly use mirrors shown as good in apt-panopticon (devuan.bio.lmu.de and devuan.packet-gain.de), both failed with the same behaviour, but also without the "byte-range" complaint, so i guess it is some other problem. I also tried tunnelling the connection through http proxy to circumvent network "problems" on provider side, but | 17:07 |
Guest9 | without effect. | 17:07 |
Guest11 | Hi! I installed Java JDK from the packages, and now I realized there is no value in the JAVA_HOME environment variable. I heard it should point to the valid JDK. Is this wrong, then? If it is, is maybe the package broken? | 17:55 |
gnarface | Guest11: nah it's just annoying, you have to set it | 18:15 |
nemo | gnarface: is update-alternatives not sufficiently flexible enough to do things like swapping out includes for the default profile? | 18:21 |
nemo | gnarface: I've never thought to wonder why JAVA_HOME was never set, have always just shrugged and manually fixed it | 18:22 |
gnarface | package maintainer's choice | 18:22 |
gnarface | some dumb ethical thing | 18:22 |
nemo | weird | 18:22 |
gnarface | you can have multiple jdk's in theory, so they want you to have to choose which java home you use manually | 18:22 |
nemo | don't get how /etc/profile.d/java_home in update-alternatives would be an ethics thing | 18:23 |
gnarface | they do it on purpose though it does seem weird | 18:23 |
nemo | gnarface: sure. but why not default it to the one the user selected as default | 18:23 |
nemo | hmmm I wonder what my gentoo does | 18:23 |
* nemo looks | 18:23 | |
nemo | $ echo $JAVA_HOME | 18:23 |
nemo | /etc/java-config-2/current-system-vm | 18:23 |
nemo | interesting | 18:23 |
nemo | I like that | 18:23 |
gnarface | i really don't know but it might have to do with not wanting to editing users' home dir configs toop | 18:23 |
nemo | it's like update-alternatives | 18:23 |
gnarface | too* | 18:23 |
nemo | but you can still override it if you want | 18:24 |
gnarface | ya but is that set globally? | 18:24 |
nemo | gnarface: well. sure. just as a default in /etc/profiles.d - you could still let users do whatever in their local | 18:24 |
gnarface | this is all about it not being global too | 18:24 |
nemo | gnarface: well it's just a default. if you're going to have a default /usr/bin/java it makes logical sense the support folder for it be global too | 18:24 |
gnarface | what i recall hearing was they figure it's not global and it might be different for every user so why not force everyone to set it manually so they have to choose and know what they chose | 18:25 |
nemo | /etc/alternatives/java /etc/alternatives/java-vm-folder | 18:25 |
nemo | something like that | 18:25 |
nemo | gnarface: then there's no reason to set /usr/bin/java symlinks either surely but. eh. whatev | 18:25 |
gnarface | i remember being really annoyed by the reasoning | 18:25 |
nemo | $ ls -l $JAVA_HOME | 18:25 |
nemo | lrwxrwxrwx 1 root root 28 Jan 4 2020 /etc/java-config-2/current-system-vm -> //usr/lib/jvm//icedtea-bin-8 | 18:25 |
nemo | nice | 18:25 |
nemo | go gentoo 😃 | 18:25 |
nemo | double // is weird | 18:26 |
nemo | script bug I guess | 18:26 |
user____ | So, <cough>, Chimaera isn't out because the manuals are being written? I just saw a possible solution... https://dilbert.com/strip/2021-09-22 | 19:34 |
Guest11 | What is the use of /etc/profile.d/ | 20:35 |
Guest11 | ? | 20:35 |
user____ | Guest11: when a new user account and directory is generated, files from profile.d are used to pre-populate the user dir. | 20:41 |
user____ | Also, default settings for various applications are stored there, when used system wide and not overridden by user settings in their dirs | 20:42 |
user____ | I see in there: bash_completion.sh gawk.csh gawk.sh vte-2.91.sh | 20:42 |
user____ | Guest11: if you read /etc/profile you see it invokes files in /etc/profile.d | 20:44 |
Guest11 | Thanks. I have vte-2.91.sh, bash_completion.sh, and vte.csh here. Why would there be a shell file "java_home"? | 20:48 |
user____ | Should set the JAVA_HOME and other vars. Look inside it. It is a text file, be sure not to edit it. | 20:50 |
user____ | It's probably installed by the JDK or JRE package | 20:51 |
user____ | You can ask the package manager to tell you what package it belongs to | 20:51 |
user____ | dpkg -S /etc/profile.d/vte-2.91.sh -> libvte-2.91-common: /etc/profile.d/vte-2.91.sh ; for example | 20:52 |
user____ | Guest11: | 20:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!