[00:04] <BraveheartBSD> An older server was depleting entropy, I had to run haveged on it, a newer server doesn't have a problem, just wondering if the entropy seed was implemented on hardware or software.
[00:16] <sdeziel> BraveheartBSD: some modern CPUs have instruction that gives fast random. Modern kernels make use of that (when available), among other sources of entropy
[00:20] <BraveheartBSD> sdeziel, thanks. I'm using an old Dell T610 as a backup, it's now working fine.
[00:29] <rbasak> BraveheartBSD: FWIW, use of haveged is questionable. See: https://lwn.net/Articles/525459/
[00:29] <sdeziel> from what I could find, this T610 uses an old CPU (Nehalem) which AFAICT, doesn't have the RDRAND instruction that more modern CPUs have
[00:30] <rbasak> Software based entropy is impossible. You have to have hardware involved somehow.
[00:30] <sdeziel> https://systemd.io/RANDOM_SEEDS/ is also an interesting quick read
[00:31] <rbasak> sdeziel: o/  while you're here, did you report a failure with ubuntu-advantage-tools and a FetchFailedException somewhere? IIRC you did, but I can't find it any more.
[00:32] <sdeziel> rbasak: no, I haven't got to reporting it yet, sorry
[00:34] <sdeziel> rbasak: doing now
[00:36] <rbasak> Thanks!
[00:43] <sdeziel> https://bugs.launchpad.net/ubuntu/+source/ubuntu-advantage-tools/+bug/2004130
[00:43] -ubottu:#ubuntu-server- Launchpad bug 2004130 in ubuntu-advantage-tools (Ubuntu) "esm-cache.service throw apt.cache.FetchFailedException when connecting to https://esm.ubuntu.com fails" [Undecided, New]
[00:44] <BraveheartBSD> rbasak, sdeziel That doesnt help, lol. The previous OS on the Dell ran fine, no entropy related errors at all. I had to change due to ZFS incompatabilty with the PERC 700. Ubuntu 22.04.1 first bought up the error, or rather, the monitoring app did.
[00:45] <BraveheartBSD> I'll keep looking into it
[00:48] <JanC> something might be using more random numbers...
[00:48] <rbasak> sdeziel: thanks!
[00:49] <rbasak> BraveheartBSD: what's the error? That your monitoring app complains? Or is there something that's actually stalling or not working?
[00:52] <BraveheartBSD> Netdata complains the entropy is depleted, and recommends running haveged, so quite possibly, netdata itself is depleteing the entropy pool.
[00:53] <BraveheartBSD> The error has no affect on the server operation.
[01:02] <sdeziel> BraveheartBSD: I think Netdata is slightly wrong/misleading here, see https://github.com/netdata/netdata/issues/2510
[01:02] -ubottu:#ubuntu-server- Issue 2510 in netdata/netdata "Warning about low entropy isn't completely correct" [Closed]
[01:04] <sdeziel> AFAIK, on 22.04 kernels, no random providing interface is blocking anymore (except a brief time window during bootup where /dev/random can block, IIRC)
[01:06] <sdeziel> that's not always been true though as /dev/random was known to block when entropy would "run low". That would then cause various issues with services needing entropy like Apache being unable to handle HTTPS connections for example.
[01:12] <rbasak> I'm curious to know whether the report of depleted entropy is resulting from just after boot, or if it's low on an ongoing basis.
[01:12] <sdeziel> https://lwn.net/Articles/884875/ is probably what I'm quoting/remembering (badly)
[01:17] <BraveheartBSD> It runs for a while, netdata displays a graph that shows entropy counting down, then posts the error. I used something different on FreeBSD 13. I'll disable netdata on this server and see what transpires.
[01:18] <BraveheartBSD> It's starting to look like Netdata is the problem
[01:54] <JanC> sounds like it might be panicking too early
[03:52] <habys> rbasak: I don't see how it's relevant. If subiquity can install one machine why not use it to install more than one machine?
[04:26] <habys> I don't need subiquity to manage any machines, I use puppet for that. I am just trying to get subiquity to update the target OS and it's crashing without any error message that I can find.
[11:48] <rbasak> habys: because you said:
[11:48] <rbasak> > I'm guessing in general people (in here|anywhere) aren't managing a fleet of physical machines imaged with ubuntu
[11:49] <rbasak> They are, but they use our dedicated solution for that.
[11:49] <rbasak> If you don't want to use the solution that everybody else uses, then you're welcome to do that but you're more likely to run into issues.
[11:49] <rbasak> But please do submit bug reports!
[12:49] <habys> thank you, I really want to get this pxe boot working
[13:20] <habys> I fixed it! It doesn't make sense but it works!
[13:22] <habys> after failing to run `apt upgrade` or equivalent in `user-data` via `late-commands:\ncurtin in-target --target=/target apt -y upgrade` and not being able to find any reason that it's not working
[13:23] <habys> I wrote a script to target with printf, chmodded and ran in with curtin and it worked
[13:24] <habys> I don't understand the crash logs in /var/crash/*.crash, they are absolutely gigantic, and don't even contain the output from the command that causes a crash
[13:57] <lotuspsychj3> !bug | habys explains filing crash bugs too
[15:59] <habys> lotuspsychj3: yes, I will file a bug, thanks
[16:12] <athos> Hi! Could someone ack LP: #2004107 so I can sync nut? I am just dropping a delta that I added a few weeks ago there that got merged in Debian :)
[16:12] -ubottu:#ubuntu-server- Launchpad bug 2004107 in nut (Ubuntu) "Sync nut 2.8.0-7 from Debian unstable" [Undecided, Triaged] https://launchpad.net/bugs/2004107
[17:01] <Odd_Bloke> Hmm, I must have missed some renewal emails for LP groups at some point, I'm no longer an Ubuntu member/developer.  How can I go about getting those restored?
[17:02] <ogra> community council should be able to re-activate you 
[17:16] <rbasak> No, it's the DMB
[17:16] <rbasak> Odd_Bloke: email devel-permissions@ please
[17:17] <rbasak> Also see: https://wiki.ubuntu.com/DeveloperMembershipBoard/KnowledgeBase#Accidental_Expiry
[18:55] <Odd_Bloke> rbasak: Ack, thank you!
[19:40] <tetrakist> Two questions. Are servers more or less efficient than space heaters at heating a room? Second, if I was given a machine with 512GB of RAM, what could I use it for?
[19:40] <arraybolt3> tetrakist: Probably not as efficient of a space heater, and with that much RAM, I'd turn it into a virtualization server.
[19:41] <bryceh> @athos taking a look
[19:41] <tetrakist> arraybolt3, why wouldn't it be as efficient? where is the electricity going, if not into heat?
[19:42] <arraybolt3> tetrakist: Processing power, I believe. This probably isn't on-topic for this room though, since this room is for support of the Ubuntu Server operating system, and not for info about server hardware.
[19:42] <tetrakist> arraybolt3, Also, I don't have anything to virtualize. Is there some place for virtualization projects?
[19:42] <tetrakist> *virtualization project ideas
[19:42] <tetrakist> arraybolt3, yeah, I'll drop the heater topic, then.
[19:44] <bryceh> @athos looks good!  yes ack on sync
[19:49] <xibalba> is there a tool on the cli i could use to highlight multiple different strings of text on the cli ?
[19:49] <xibalba> i know less can do it for 1 string, but i have 3 different strings i'd like to highlight
[19:50] <xibalba> as a workaround i could do tmux, split the screen into 3 windows and 1 highlight per window
[19:51] <sdeziel> xibalba: with less, you can search for a regex like "/(foo|bar)" and all occurrences of `foo` and `bar` would be highlighted
[19:51] <xibalba> brilliant
[19:51] <ravage> echo "pattern1 pattern2 pattern3 pattern4" | grep --color -E '^|pattern1|pattern2'
[19:51] <xibalba> i need to see some lines above/below the strings
[19:52] <xibalba> but i dont know how many linea above/below, so -A -B options in grep wont work
[19:52] <sdeziel> xibalba: with grep, you can show the surrounding 3 lines with `grep -3`
[19:52] <sdeziel> ah, sorry replied too quickly
[19:52] <xibalba> no worries the less option just worked perfect, thank you sdeziel 
[19:52] <xibalba> thank you too ravage 
[19:52] <sdeziel> cool
[20:36] <athos> thanks, bryceh :)
[20:39] <athos> sync'd