[00:58] <Mr_Cyclops> Hello. Is it possible to have Google Photos on Ubuntu? I mean, run an application or webserver, add the folder that has all the pics, and get the view with face taggin etc. like it works on google photos? thank you
[01:20] <pragmaticenigma> Mr_Cyclops: What you described, already exists as Google Photos. Google has not released any applications for any computer to interface with that web application. As far as homebrew, start searching the web. That is all relatively new territory and I haven't heard of any projects for home users that are ready for prime time
[01:20] <Mr_Cyclops> pragmaticenigma, :) It restricts the usage to 15GB, my collection is in TB ...
[01:21] <Mr_Cyclops> when I mean restricts, I mean, 15GB is free ... :)
[01:21] <Mr_Cyclops> which you know of course
[01:21] <Mr_Cyclops> :d
[01:22] <pragmaticenigma> Mr_Cyclops: that sounds like a you problem... they have paid platforms for more storage
[01:22] <Mr_Cyclops> pragmaticenigma, yup, I can't afford it :)
[01:22] <Mr_Cyclops> but thanks anyway ... appreciate it
[01:23] <pragmaticenigma> Mr_Cyclops: If you can't afford a paid platform, then I would say that you probably can't afford all your photos to be corrupted because of a configuration mistake either
[04:31] <lotuspsychje> good morning to all
[06:34] <lordievader> Good morning
[07:00] <Sveta> !indonesia
[10:59] <marcoagpinto> Heya!
[10:59] <marcoagpinto> >:)
[11:32] <akem-hp> Hey.
[11:33] <akem-hp> Got myself a new 4TB external drive for backup, time for file copying - long time ;)
[11:35] <lotuspsychje> wich brand
[11:35] <akem-hp> Seagate, it was relativly cheap, 90 euros for 4 TB.
[11:35] <lotuspsychje> sweet deal
[11:36] <lotuspsychje> im also a seagate fan
[11:36] <tomreyn> that'll be shingled, i guess
[11:37] <akem-hp> I don't which one bought the other, but some drives are double branded Seagate/Maxtor nowdays.
[11:37] <akem-hp> But this one is clearly labeled Seagate everywhere on the packaging and on the drive.
[11:51] <TJ-> akem-hp: check which type the drive is before writing anything to it - if it is a Seagate "Archive" drive it uses Shingled Magnetic Recording (SMR) and there are two versions of those - drive-managed and host-managed. Later devices are host-managed and so you may need to ensure operating system support for it
[11:53] <akem-hp> TJ? i already started copying to it, i mounted it, it was formated already, not sure what type it is but it looks like its copying fine atm, TJ how can i tell the type of drive?
[11:54] <akem-hp> At about 100 MB/s: Copying 2.36 TiB   4% |##                                                       | 106.28 MB/s ETA:  6:28:11
[11:55] <lordievader> lotuspsychje: Seagate is still leading in drive failures in backblaze's hard drive failures report: https://www.backblaze.com/blog/backblaze-hard-drive-stats-q1-2019/
[11:55] <lordievader> akem-hp: `smartctl -a /dev/sdX` typically tells you the make and model.
[11:56] <lordievader> Unless behind a raid controller or something.
[11:57] <TJ-> akem-hp: "grep . /sys/block/sd?/device/model"
[11:58] <TJ-> akem-hp: SMR drives use overlapping tracks grouped into zones. When a write is done to a track the drive has to re-write every track in the zone
[11:59] <TJ-> akem-hp: to prevent problems with very low write throughput and other issues the drive (or the Host OS) needs to be aware of the zones and their size and handle them correctly
[12:00] <akem-hp> https://pastebin.com/5iRf3Vd7
[12:01] <TJ-> akem-hp: arghh, behind a USB bridge
[12:01] <akem-hp> It's plugged on my USB hub.
[12:01] <TJ-> akem-hp: I'd strongly recommend you determine the exact model ID and firmware version from the device's sticker in that case
[12:02] <akem-hp> I can stop the copying and plug it directly to the computer.
[12:02] <TJ-> akem-hp: from those you can search for the seagate specs that'll tell you if it is SMR, and if so whether it is drive- or host-managed
[12:03] <TJ-> akem-hp: if the device is host-managed you'll need to use the dm-zoned device-mapper layer. See https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/device-mapper/dm-zoned.txt?id=3a564bb3a8a6950e18b1f5d209bda39fc3831074
[12:06] <marcoagpinto> ahhh.. I too bought a WD 4 TB HDD months ago to convert all my DVD films to .iso
[12:06] <marcoagpinto> :)
[12:06] <marcoagpinto> but for months that I don't do it
[12:08] <tomreyn> TJ-: this is the first time i hear you need dm-zoned for SMR. is this documented anywhere in ubuntuland, to your knowledge?
[12:08] <tomreyn> host-managed SMR, that is.
[12:12] <tomreyn> there's also https://lwn.net/Articles/720226/
[12:13] <tomreyn> mkfs.ext4 on 18.04 does mention two *lazy* options.
[12:17] <akem-hp> TJ-, Well i plugged it directly to the computer, i got the smae outputs for both commands - this is dmesg https://pastebin.com/C6gZSVX4, and on the papers or stickers there is nothing like that, just serial numbers but i cannot get info with them online, it's just written Expansion, not archive...But this is for backup, it won't be plugged very often.
[12:19] <TJ-> tomreyn: it's mainly an Enterprise issue for read-mostly SANs so there's not much info for consumers
[12:20] <TJ-> akem-hp: does the drive have a native USB3 interface then?
[12:20] <tomreyn> enterprises using SMR disks?
[12:20] <akem-hp> TJ-, yes.
[12:21] <akem-hp> Hopefully, i wouldn't get lower than 3.0 for large data copying like that, it would be too slow.
[12:22] <TJ-> akem-hp: looks like the STEF4000400 (Expansion+ 4TB)
[12:23] <tomreyn> akem-hp: does hdparm -i return anything more useful?
[12:23] <TJ-> tomreyn: yes, for read-mostly online storage (that's why they're called "archive" drives)
[12:24] <tomreyn> i imagine it'd make sense for aws glacier or the like
[12:24] <akem-hp> tomreyn, No: SG_IO: bad/missing sense data, sb[]:  70 00 05 00 00 00 00 0a 00 00 00 00 24 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 HDIO_GET_IDENTITY failed: Invalid argument
[12:25] <tomreyn> akem-hp: how about   hdparm -I   (upper case)
[12:26] <tomreyn> oh you need sudo for both
[12:26] <akem-hp> Yes i used sudo, there is more infos there with -I.
[12:26] <tomreyn> but you wont share it? :)
[12:27] <akem-hp> But i don't see anything related to SMR or something, wait i'm pasting it. :)
[12:27] <akem-hp> https://pastebin.com/HTGUX8yH
[12:28] <tomreyn> from what i read online, those drives marketed as Seagate Expansion Desktop can be all kinds of models
[12:28] <TJ-> akem-hp: you  might be OK, it looks like there are too model-streams "Expansion" and "Backup". 'Backup' are definitely SMR but cannot find any specs for the "Expansion" stream
[12:29] <akem-hp> TJ-, I see, it's good to know about it, i didn't know about SMR at all, i'll pay attention to this when buying harddrives from now on.
[12:30] <TJ-> hmm, that hdparm output is hilarious: "Nominal Media Rotation Rate: 29295"
[12:30] <tomreyn> thanks for sharing, akem-hp. the drive identification looks like garbage to me.
[12:31] <TJ-> the only info (which doesn't make it clear) is here, but it calls the 4TB "Expansion+" https://www.seagate.com/gb/en/consumer/backup/expansion-portable/#specs
[12:31] <akem-hp> tomreyn, Yes, it's just exactly like that in the terminal too. :/
[12:32] <TJ-> tomreyn: akem-hp  garbage for sure, due to the USB bridge :) I hate how the drive makers are hiding the devices behind the USB translation
[12:32] <tomreyn> TJ-: seems to depend on the page language, german shows both the Expansion+ STEF4000400 and Expansion STEA4000400   4 TB models
[12:33] <akem-hp> TJ-, the last one 4TB on the link you posted is Expansion without "+" it's the same i think.
[12:33] <akem-hp> Yes i'm looking at the french version...
[12:33] <tomreyn> the UK page variant only knows about the + one
[12:34] <tomreyn> and the pdf file names even have 'emea' and the country code in them
[12:34] <tomreyn> geez
[12:34] <TJ-> apparently the only difference between "Expansion" and "Expansion+" is the latter is shipped with some Seagate drive management software
[12:35] <BluesKaj> Hi folks
[12:35] <marcoagpinto> BluesKaj!!!! Hello!!!!
[12:35] <marcoagpinto> >:)
[12:35] <akem-hp> Hey.
[12:35] <BluesKaj> hi marcoagpinto
[12:35] <marcoagpinto> I am drinking some cola
[12:35] <marcoagpinto> :)
[12:37] <BluesKaj> welk, it's your pancreas...
[12:39] <tomreyn> BluesKaj!!!! Hello!!!!
[12:39] <tomreyn> I am drinking some tea
[12:40] <BluesKaj> hey tomreyn
[12:40] <tomreyn> just so you know!!!
[12:40] <BluesKaj> hehe
[12:42] <akem-hp> A bit offtopic but Termux is great on Android without need for root.
[12:43] <TJ-> akem-hp: I'd presume as the drive is behind a USB bridge, if it is SMR, it has to be drive-managed, so the only effect you might see is a slow-down in sustained write throughput once it fills its internal cache RAM
[12:44] <BluesKaj> the caffeine probly doesn't hurt your organs much, but all that sugar isn't good marcoagpinto
[12:45] <akem-hp> TJ-, ok, i see, but i'll use it for very occasional writes as it is backup unit, actually i'm writing to it full speed.
[12:48] <lordievader> akem-hp: Termux is lovely
[12:48] <lordievader> Through a termux script I can download youtube videos by sharing the link from youtube with termux.
[12:48] <akem-hp> lordievader, yep :) it has apt just like Ubuntu.
[12:49] <akem-hp> Didn't know you could do things like that.
[12:50] <lordievader> akem-hp: https://medium.com/@baradhiren07/download-youtube-videos-using-termux-2e72082d5173
[12:50] <akem-hp> I tried to run my small ruby IRC bot on it, it worked just fine :P
[12:53] <akem-hp> lordievader, nice.
[13:28] <tomreyn> 3 days left for 18.10
[13:32] <BluesKaj> good riddance...it was buggy in my experience, and some of the bugs still show up in the newer releases
[13:33] <BluesKaj> one of them is a dev release, but the annoyances are still there
[13:54] <tomreyn> blackflow: i like that you have an opinion and that you state it. i prefer you not calling 'BS' when others voice theirs (note how i did neither).
[13:57] <blackflow> I call it as I see it. Feel free to call my opinion BS if you think it is.
[13:57] <tomreyn> regarding 'generic', OP asked for "the recomended options", which i rephrased (translated?) as a "general recommendation".
[13:58] <blackflow> (in particular for LVM snapshotting being a "traditional" route, what with all its bugs and performance issue, it's anything _but_ "the" traditional route)
[13:58] <tomreyn> it's what's been supported for a long time, and what's been used for a long time.
[14:03] <tomreyn> i'm not saying it's the technically greatest solution, it certainly isn't. i'm not saying there are no bugs there. but lvm snapshotting is tried and tested, the deficits are well understood. for people who prefer a somewhat conservative approach, it can be a good choice. and with data, I want conservative, and i think many others want it, too.
[14:04] <tomreyn> i'm by no means saying your approach is wrong, i'm saying that it's probably not the solution which works for the average user.
[14:06] <blackflow> They asked for recommended. I recommended based on tried and tested, personal experience, which I'm also dogfooding. If they don't want it or don't feel confident about ZFS, fine. It's still my recommendation. And not just mine.
[14:08] <tomreyn> yes, but you didn't state so, so i listed other approaches.
[14:08] <blackflow> but shooting it down just because you noobs can't use anything other than what ubiquity is pre-installing for you.... well.... hilarious.
[14:08] <tomreyn> now you call me a noob, nice. are we getting personal now?
[14:09] <tomreyn> maybe you should take a break and cool off a little.
[14:11] <BluesKaj> I guess if one uses a conservative approach it's considered noob, by some more advanced users
[14:11] <blackflow> You know what, you're right. I'm sick of this crap anyway.
[14:12] <BluesKaj> angry young man attitude
[14:15] <tomreyn> he certainly knows a lot more about zfs than some (or many) of us, definitely me - no doubt there. i'd just hope for a better environment for discussions, and one where we can point out multiple solutions to solving stated problems / tasks.
[14:17] <RikMills> o_O
[14:22] <lordcirth> tomreyn, was the snapshot discussion for a root partition?
[14:23] <tomreyn> lordcirth: all we had to work with was: <JustJohnny> what are the recomended options to implement bare metal backup in CLI-only ubuntu server?
[14:25] <lordcirth> Ah, that's not much.
[14:26] <TJ-> I don't think "works for me" is a good basis for a recommendation, when considering https://github.com/zfsonlinux/zfs/issues?q=is%3Aissue+is%3Aopen+label%3A%22Type%3A+Defect%22
[14:27] <TJ-> What with those and the recent performance issues due to the 5.x kernel changes affecting SIMD/FPU access as I understand it
[14:30] <lordcirth> The performance issue is a concern. The others would not prevent me from running 0.7.x in prod. 0.8.x is another matter.
[14:31] <lordcirth> Unfortunately there's still nothing that really competes with ZFS. Btrfs is ok for rootfs's, since ZFS on root is annoying to set up.
[14:32] <TJ-> I think one other issue is always going to a concern. Learning-curve, experience, testing, before using it in anger
[14:32] <lordcirth> Learning is certainly needed, yes. But that's true of everything, really.
[14:32] <TJ-> ZFS hasn't grown support for zoned devices has it (ZAC/ZBC) ?
[14:33] <TJ-> lordcirth: right... but someone familiar with general Linux admin is going to be better off with bacula/rsync type back-up solutions than block-device/file-system specific solutions, I think
[14:34] <lordcirth> I don't think it supports zoned devices, no. Are they common? I've never had one.
[14:34] <lordcirth> Yes, for the question as phrased I would have replied "BorgBackup"
[14:37] <TJ-> host-managed SMRs are becoming quite common because they're cheap. we can use dm-zoned device-mapper to handle them with anything on top of course, so that could be used under ZFS as with other options like LVM, but then the purported advantage of ZFS foing everything is lost
[14:37] <TJ-> err s/foing/doing/
[14:40] <lordcirth> Well, the main advantages of ZFS are checksumming, redundancy / failure handling, and snapshots. If using dm-zoned underneath works well, I'd use it. But I suspect there'd be a performance hit in practice.
[14:41] <TJ-> Less than not using it :)
[14:41] <lordcirth> But then, if you are using SMR you have already traded away performance
[14:41] <TJ-> SMR is fine when used with host-aware OS tooling (or device-aware)
[14:42] <TJ-> the problem is when using a host-aware drive and no OS support
[14:48] <lordcirth> TJ-, what performance do you get with proper OS support?
[14:49] <TJ-> lordcirth: the zones are mapped so there's a metadata zone hidden from 'users' and then user zones for data, which prevents constant re-writing of entire zones just due to metadata changes
[14:50] <lordcirth> And the metadata zone has different performance? Or is it just so that you can rewrite just that zone?
[14:50] <TJ-> see https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/device-mapper/dm-zoned.txt
[14:50] <lordcirth> It's .rst now, apparently
[14:52] <TJ-> well doh! teach me to delete the original commit ID from the URL for clarity!
[14:57] <tomreyn> "dmzadm" is not an ideal choice for a name.
[14:58] <TJ-> yeah, sounds like a networking tool
[14:58] <tomreyn> there are too many acronyms in IT ;)
[14:59] <TJ-> Maybe its for use on the 38th parallel (north/south korea border)
[14:59] <TJ-> there's only two in IT, I and T :)
[15:03] <tomreyn> :-)
[19:22] <coffeecow> lotuspsychje,
[19:22] <lotuspsychje> welcome coffeecow
[19:22] <lotuspsychje> you can launch a poll or discussion about ubuntu here fi you like
[19:22] <lotuspsychje> *if
[19:26] <coffeecow> how do I launch a poll, and thanks lotuspsychje
[19:26] <coffeecow> I was asking "Is there anyone here who prefers Seahorse over KeePassXC after trying them both?"
[19:26] <lotuspsychje> im fan of using none myself coffeecow
[19:27] <coffeecow> I keep hearing bad stuff abotu seahorse but I'm not sure if it really is any less secure. All I know is it integrates like a thousand times better than keepassxc does.
[19:27] <coffeecow> Yeah but I like using different passwords for *everything*.
[19:27] <coffeecow> My memory ain't that good.
[19:28] <coffeecow> I don't need someone to be able to build a profile of me from like... a database leak a la haveibeenpwned.com
[19:29] <lotuspsychje> coffeecow: leaked password from that sorts of sites, usualy come from owned websites
[19:29] <lotuspsychje> coffeecow: the security part of ubuntu, is keeping your system up to date, and often change password, use complex passwords
[19:30] <lotuspsychje> maybe other volunteers can point you do the pro & contra of seahorse and keepassx
[19:30] <daftykins> KeePassXC is a service that lets you maintain a local database file, so there's no cloud involvement
[19:31] <daftykins> been looking at that to move clients to, since they're so terrible with passwords + security
[19:35] <lotuspsychje> https://www.forbes.com/sites/jasonevangelho/2019/07/12/ubuntu-executes-another-massive-change-to-the-way-it-updates-proprietary-nvidia-drivers/#1850b6597aa8
[19:43] <TJ-> !info pass | coffeecow: gpg + pass + git
[21:12] <jeremy31> Wait for it....
[21:17] <OerHeks> 1.. 2 .. ?
[21:18] <jeremy31> UWN
[21:35] <Bashing-om> UWN587 is on the streets: https://wiki.ubuntu.com/UbuntuWeeklyNewsletter/Issue587 :D
[21:53] <OerHeks> yeah, good news for Zen 2 users https://www.phoronix.com/scan.php?page=news_item&px=AMD-Releases-Linux-Zen2-Fix
[22:04] <TJ-> OerHeks: thought you were on about my UK ISP, Zen!
[22:23] <TJ-> phew! looks like we solved the apache2/openssl 1.1 DoS
[22:24] <TJ-> Had my head deep inside libssl1.1/apache2 debugging with gdb most of the evening
[22:25] <sarnold> TJ-: what was it?
[22:25] <TJ-> SSL_error queues being mis-handled and the wrong errors causing 2 processes to spin at 100%
[22:26] <TJ-> bug #1836329
[22:27] <TJ-> looks like it needed two commits cherry-picking to solve it
[22:28] <tomreyn> thanks for working this out.
[22:29] <TJ-> I wasted 1/2 hour before I realised there are no -dbgsym packages for apache :s
[22:30] <TJ-> well, not for the version in 18.04 anyhow; not sure why that is, I thought there was a debhelper that generated those automatically
[22:30] <tomreyn> there seem to be a couple packages without ddebs
[22:31] <daftykins> when does this situation arise, is it affecting anyone running apache on bionic?
[22:31] <tomreyn> not sure why either.
[22:32] <tomreyn> looks like #1836329 should be public security really
[22:35] <TJ-> daftykins: almost any client could cause it
[22:36] <TJ-> daftykins: it seems to be 2 scenarios we know of, 1) client protocol renegotiation and 2) client X509 certificates
[22:37] <daftykins> hmm doesn't seem like a good situation