Posts by grawity@social.treehouse.systems
(DIR) Post #B61uFD4gVqHggrvv3Q by grawity@social.treehouse.systems
0 likes, 0 repeats
well okay I guess
(DIR) Post #B67gzymXEmKfYxbvAe by grawity@social.treehouse.systems
1 likes, 0 repeats
@domi I've done this via separate tools (writevt and peekvc) but didn't know there was a combined version
(DIR) Post #B67ks3FZ8lMmv9fGq0 by grawity@social.treehouse.systems
0 likes, 0 repeats
@domi oddly I don't have it in my busybox, I had to install it as a separate package (https://aur.archlinux.org/packages/conspy)
(DIR) Post #B6PndXLmk7Zodhcga0 by grawity@social.treehouse.systems
1 likes, 0 repeats
"root@server1# reboot"ok, while it's rebooting, let's check server2...wait, why is server2 powered off? it was still running a minute ago?oh, right, server1 has the UPS USB cable, and when NUT ups-monitor on server2 lost communications with NUT upsd on server1 (due to server1 rebooting), it somehow decided that's a critical battery event and immediately powered the server down
(DIR) Post #B6PquXoKaGn8qvmCRs by grawity@social.treehouse.systems
1 likes, 0 repeats
when rebooting servers, it's really reassuring to have serial console access (or at least graphical KVM) and to be able to watch a server boot, all the way from GRUB if not necessarily the BIOSfeels like it should be a baseline feature on every PC and every VM, but unfortunately isn'tI have a ~/bin/rcons which is just a big case..esac block which takes a hostname and does whatever is necessary to reach the remote console of that device, whether it's "ipmitool sol" or "telnet cisco 2005" or "~/bin/ociconsole" or "echo Go to the Hetzner website"
(DIR) Post #B6gPOP3werWsk4bASu by grawity@social.treehouse.systems
1 likes, 0 repeats
I'm also a bit confused as to why my Comet lake laptop has been feeling just as sluggish as the previous Haswell one. It isn't even showing any high CPU load or anything, it isn't set to powersave mode in GNOME, but even scrolling in Firefox feels a bit janky at times. YouTube videos drop frames. Takes half a minute to resume from sleep. Wonder if it's an iGPU driver thing
(DIR) Post #B6kNynGMjjAowmTLYO by grawity@social.treehouse.systems
0 likes, 0 repeats
Work mail server kept rejecting mail due to rDNS lookup failures, and was logging request timeouts internally... and sure, rDNS + DNSSEC validation = slow, but it seems like the Unbound resolver sometimes just drops queries on the floor? More than once, tcpdump showed me a query arrive at the resolver and nothing go out, not even a servfail error.I assume that *has to be* caused by something trivially tunable, like increasing Unbound's receive UDP buffer size maybe – but I decided to just switch the resolver to BIND instead, and so far it's been running with zero complaints with the default knobs &shrug;(...well, except for having to raise a "max records per type" limit, because a local hosting company does their SPF by pointing "a:" to a subdomain which has 190 A records listing every individual IP address of theirs, instead of doing "ip4:<cidr>" like a normal)
(DIR) Post #B6ssnYCb2knEtYVXHM by grawity@social.treehouse.systems
1 likes, 0 repeats
scratch that, today's "oh come the fuck on this is stupid" in apparmor is *they made /usr/bin/tshark no longer allowed to read pcap files*
(DIR) Post #B7O1piCkmQl6D3FXtI by grawity@social.treehouse.systems
1 likes, 0 repeats
@jana yeah this was originally created for a "shell account" service that had ~a dozen servers and everyone was on IRC and/or on SSH with the hostnames visible; back in 2012 it was still "fine" to reveal it by default... but it does have a mode to hide all of the "source" hostnames/addresses, at least.(it actually sits behind a "finger reverse proxy" which specifically handles things like crawler GET requests and forwards them directly to iptables, etc., etc., while forwarding everything else to the actual rwho fingerd)one thing I've noticed missing from your implementation is support for "finger -l", i.e. "/W" query prefix, it should at least be handled and trimmed even if it doesn't change the output
(DIR) Post #B7YNLFcbEQeWM5hK1w by grawity@social.treehouse.systems
1 likes, 0 repeats
smartphones: wi-fi address randomization to become harder to trackmeanwhile: so many cars trackable by their "Audi_MMI_" or similar Wi-fi, or dashcams with their "70mai_" SSIDs
(DIR) Post #B7ggywF2pD5xmBMMxU by grawity@social.treehouse.systems
1 likes, 0 repeats
It's pretty handy that I can connect an external HDD to my wifi router, do "/disk/set 0 nvme-tcp-export=yes", and then "sudo nvme connect" to it from my home server. the sole USB port on the router is 2.0, unfortunately, but it might do (at least until I figure out the power socket mess and move the USB HDD directly to where the server is...)
(DIR) Post #B7pSEGaHiwwIJzmjiq by grawity@social.treehouse.systems
0 likes, 0 repeats
@domi I'm now curious about how the -p option in your `wall` is implemented
(DIR) Post #B7phpYgwrWaAVY8sJE by grawity@social.treehouse.systems
0 likes, 0 repeats
@la @domi does that mean you can have a polyglot that is both ISO BMFF and ISO 9660?(the latter also reserves 16 sectors &times; 2048 bytes for "system" use)(which AFAIK originally meant "DEC VAX boot sector")
(DIR) Post #B7phpYz1mJ2nPcbJUO by grawity@social.treehouse.systems
1 likes, 0 repeats
@la @domi that reminds me of the mess that is "isohybrid" "dd-able to USB stick" .iso images, which are not only four-way polyglots (MBR, GPT, Mac APM, ISO-9660) but also have different contents depending on what *media* they're written on:- if you write the .iso to a CD (with a logical block size of 2048B), you find an ISO-9660 Volume Descriptor Set at sector 16, and an Apple Partition Map with two entries at sectors 1-2- if you write the .iso to a USB stick, you instead find a GPT partition table at sectors 1 and 12..63, whereas sector 16 is now just an empty GPT slot (because a USB stick's logical block size is now 512B)- the GPT header is deliberately configured to start the actual "table" from sector 12 so that it jumps over the two APM entries (which are now in sectors 4-7 and 8-11 when viewed through a 512B lens)- sector 0, of course contains a MBR with x86 boot code when booting from USB stick (ignored when booting from CD as the first 16 sectors are "reserved system area")
(DIR) Post #B7vKLlScJpPNlyS7m4 by grawity@social.treehouse.systems
0 likes, 0 repeats
# networkctl reloadFailed to issue io.systemd.service.Reload() varlink call: No such device or addresssure would prefer if there had been some transition period where the client tools could fall back to the D-Bus calls...
(DIR) Post #B7vKLlg5VkBSRkksls by grawity@social.treehouse.systems
0 likes, 0 repeats
the version of networkd that was running was v260, the version of networkctl that I'm now running is v261, don't tell me that it's unreasonable to maintain backwards compat across 2 adjacent releases at the very least
(DIR) Post #B7vMACV89mSu4SF40e by grawity@social.treehouse.systems
0 likes, 0 repeats
@astraleureka sluggish hog? can't say I've seen dbus-daemon acting like that, usually it does not see enough traffic to be a botherVarlink is, well. on the one hand, it's just JSON over a Unix socket (almost like JSON-RPC but incompatible) – so it's traditional direct peer-to-peer IPC, not a bus or anything like thaton the other hand, it's JSON. and apparently with some ideas from gRPC, since it's no longer a 1:1 call-response (and deliberately not JSON-RPC compatible), there can be multiple replies per call (streaming), etc.it also has protocol switching, apparently systemd now has a Varlink service where you send a specific call, and the varlink JSON socket becomes an SFTP socket.so it's more efficient than d-bus, doesn't require a daemon, you just have a bunch of sockets in /run... but for me personally it feels like much more of a mess than d-bus ever was. like, it has a kind of docker/k8s "HTTP over unix socket as a local IPC mechnanism" vibe. don't ask
(DIR) Post #B7vOzw0iueM3tRodMG by grawity@social.treehouse.systems
0 likes, 0 repeats
@astraleureka yeah I'd say that's an overuse of dbus, surely there was a better way to deliver button presses. Like rely on the shared bus as a discovery point, but have a dbus call to get a direct unix socket FD and then use that for low-latency traffic(even regular GNOME/KDE/etc. use a separate bus instance for accessibility events...)I was only kinda following the kdbus story from the systemd side, and IIRC some of the negative response to it (when it started sprawling into things like bloom filters) was that allegedly *much* of the performance gain was not due to context switches etc., but due to dbus-daemon being slow? so I wonder if dbus-broker generally performs any better than the older dbus-daemon.(there was apparently a 2nd try at kdbus – "bus1" – but I've not followed that at all, in part because most of the people involved had moved out of IRC)
(DIR) Post #B80OTnGH13eBC8Obh2 by grawity@social.treehouse.systems
0 likes, 1 repeats
@domi if there ever was a situation to joke about how instead of "(resilvering)" this should say "Verifying dmi Pool Data..."
(DIR) Post #B81jVB9lx6f5mzoi1I by grawity@social.treehouse.systems
0 likes, 0 repeats
@ariadne the NT 5.x code is still up on github so I wouldn’t be surprised it got ingested