**** BEGIN LOGGING AT Mon Jun 12 03:00:02 2017 Jun 12 07:37:03 DocScrutinizer05: thanks. i wonder if i'll really manage. i don't really have the dispositon. will try my best though Jun 12 07:37:50 facilitate plz Jun 12 07:51:40 DocScrutinizer05, thank you. I'll do my best to keep the confidence. Jun 12 12:16:26 L29Ah_: shut up plz Jun 12 12:16:44 or grow a brain Jun 12 13:32:15 i'll consider your proposition thoroughly Jun 12 14:19:06 it's up to you to refuse to understand what service the council does voluntarily for benefit of the maemo community, but please don't expose your ignorance by posting disrespectful/silly comments related to council Jun 12 14:20:52 i wonder how efficient is N900's handover between cells, esp. 2G<->3G. Jun 12 14:21:07 'normal' Jun 12 14:21:17 afaict Jun 12 14:21:21 i still get 403 on maemo.org so fuck council and community Jun 12 14:21:22 i have very bad experience when using my cellular net, but the SGS4 has almsot no issues Jun 12 14:22:38 sicelo: facilitate plz Jun 12 14:22:40 "fuck community" in maemo community channel - a nice idea Jun 12 14:23:07 L29Ah_: stop that bullshit or earn a ban! Jun 12 14:23:18 L29Ah_: i'm not council. :-) Jun 12 14:23:52 you'll get my vote if you promise -o DocScrutinizer05 and fix 403 on maemo.org :] Jun 12 14:24:12 last warning Jun 12 14:25:10 DocScrutinizer05: seriously, where can i see the list of achievements of the council or techstaff for, say, the last year? Jun 12 14:25:53 you wouldn't even understand if I stub your nose into it Jun 12 14:26:34 council is not even supposed to thrive for "achievements" Jun 12 14:26:49 neither are they in charge of tmo administration Jun 12 14:27:18 what are they supposed to do except spamming me with some stupid polls regarding people and packages? Jun 12 14:27:19 s/thrive/strive Jun 12 14:28:08 and how do i access maemo.org if techstaff is effectively dead? Jun 12 14:28:31 so community, much bullshit Jun 12 14:28:32 you might appeal them when I +b you, so they could ask me to revert that ban Jun 12 14:28:45 they're not dead, just .... not always available :) Jun 12 14:28:57 i guess Jun 12 14:28:59 DocScrutinizer05: I don't see the point of banning people for that Jun 12 14:29:27 bencoh: L29Ah_ has a long record for trolling against council Jun 12 14:29:46 "so be it"? it's not like he's spamming Jun 12 14:29:50 long as in 2years" Jun 12 14:30:50 bans are also earned by spoiling channel mood aka rising temperature, not only for spamming Jun 12 14:31:25 trolls usually don't spam, they troll Jun 12 14:31:40 just take it easy already Jun 12 14:31:51 right, not worth it Jun 12 14:31:56 exactly Jun 12 14:32:03 sorry Jun 12 14:32:07 cranky mood Jun 12 14:32:11 :) Jun 12 14:33:05 ~techstaff Jun 12 14:33:05 methinks techstaff is techstaff(AT)maemo.org - the folks that keep your maemo infra running. Devotion to Duty http://xkcd.com/705/ Jun 12 14:33:13 techstaff supposed to neot be "dead", send email or ping sysops or admins in here directly Jun 12 14:33:16 got disconnected Jun 12 14:33:42 i did Jun 12 14:33:51 sicelo: manual 2G->3G transition usually works here, with no data disconnection Jun 12 14:33:54 particularly blacklist on tmo is chem|st's domain entirely Jun 12 14:34:15 council not involved in that at all Jun 12 14:34:29 neither is "community" Jun 12 14:34:54 I hate to point at people, but iirc some of the techstaff people hang around in here (xes maybe?) Jun 12 14:35:19 so if you don't get any answer by mail, maybe pinging here might help Jun 12 14:35:19 i wonder if i have h/w problem Jun 12 14:35:34 they were both highlighted the last time i whined about the 403 stuff Jun 12 14:35:43 xes__ warfare Jun 12 14:35:47 chem|st Jun 12 14:38:48 of i select 3g only, then eventually i have no network access at all (Allbars disapper). Jun 12 14:39:38 sicelo: this could be network-related Jun 12 14:40:00 the way blacklist work in talk.maemo.org is quite... suboptimal. I "whined" about that myself a lot. Only meaningful approach though is to discuss the issue with the responsible tmo admin ( chem|st) and offer to help improving it Jun 12 14:40:36 chem|st: okay let's discuss it! Jun 12 14:41:01 yes i thought so bencoh for a long time. however the s4 has np issue at all Jun 12 14:42:03 the rationale I heard was: "we can't limit blacklist to login-only since the rogue IPs run DOS attacks against the server, so we need to block mere reading access too." I have a hard time undertsnading how a blacklisting with a 403 page leverages DOS attacks Jun 12 14:42:37 understanding* Jun 12 14:43:31 I'd suggest to rather employ fail2ban for such DOS-defeat Jun 12 14:44:14 i wonder who cares about maemo enough to do DoS attacks Jun 12 14:44:33 me too Jun 12 14:44:34 russian hackers i presume Jun 12 14:44:41 ™ Jun 12 14:45:19 anyway to discuss the issue chem|st and maybe xes__ are the ones to ping Jun 12 14:45:44 DocScrutinizer05: well, 403 helps if DoS involves triggering lots of dynamic content generation Jun 12 14:45:55 :nod: Jun 12 14:46:29 so employ fail2ban to block unregistered users that trigger lots of high CPU load Jun 12 14:47:01 my humble 2 cent Jun 12 14:47:54 in a lot of forum software I see e.g. "search" protected by capcha for anonymous access Jun 12 14:48:42 can't be impossible to get tmo on par with the bazillion other fora out there Jun 12 14:50:34 chem|st first and foremost is tmo chief moderator, the technical server administration (and thus parts of the blacklisting and DoS-defense) could get assigned to a co-admin Jun 12 14:50:53 THIS would be a job for council, to decide Jun 12 14:51:56 disclaimer: my info is like 3 years old, don't know if it's still up to date Jun 12 14:52:54 for your question about "achievements of council": council is supposed to publish a meeting-minutes for each weekly meeting Jun 12 14:53:50 should be both on wiki/planet/tmo whatever, and in ML Jun 12 14:54:13 DocScrutinizer05: blacklisting and DoS is techstaff's domain Jun 12 14:54:31 not mine Jun 12 14:54:44 ok, sorry Jun 12 14:54:49 :-) Jun 12 14:54:52 np Jun 12 14:54:56 so xes and warfare? Jun 12 14:55:03 mainly xes Jun 12 14:55:08 ta! Jun 12 14:55:16 L29Ah_: ^^^ Jun 12 14:55:45 the problem with the spammers is that it is "human farming", they bypass like everything we throw at them Jun 12 14:55:59 o.O Jun 12 14:56:20 fail2ban / throttle, then? Jun 12 14:57:01 "sorry, your usage pattern triggered some safeguard. Please log in again after the 15min grace period expired" Jun 12 14:57:02 not even that will work, we train them to get smarter ourselves... Jun 12 14:57:30 it is a single IP managing to create an idle account on first attempt Jun 12 14:57:58 that account is most probably sold away as most of them wait at least a 30day grace period Jun 12 14:58:08 get quota for users Jun 12 14:58:10 some spamposts get edited too Jun 12 14:58:27 quota? they do a single post Jun 12 14:58:38 CPU load quota Jun 12 14:58:52 not talking about SPAM Jun 12 14:59:00 DoS Jun 12 14:59:11 that is a different issue and not the server but the IPHH backbone is the target Jun 12 14:59:23 the issue with SPAM is known and annoying and difficult to defeat Jun 12 14:59:49 duh! IPHH? Jun 12 15:00:04 what did they do to make skiddies angry? Jun 12 15:01:00 it is a backbone they could use, so they try to get inside servers, that is managed well iirc, the account creation attempts on garage and tmo are the left issues Jun 12 15:01:28 trying to do php injections to hijack the servers and stuff Jun 12 15:02:14 convert account creation to something annoyingly slow, like email with manual approval? Jun 12 15:02:50 and that happens mostly from ranges we have in our blacklists, most of the time noone gets hurt, but every now and then ISPs change areas so we see IPbans on legit networks as it takes a while to be removed from such blacklists, we update daily though Jun 12 15:03:27 IP bans always are problematic Jun 12 15:03:57 nobody found a better alternative though Jun 12 15:04:07 not yet Jun 12 15:04:13 except changing the user interface, like suggested before Jun 12 15:04:22 xes tried all kinds of things Jun 12 15:05:27 we changed the UI... not visually, but the inner workings to prevent bots, that is why we guess it is human farms Jun 12 15:06:01 chem|st: could the forum software learn to show a certain class of posts (those from non-yet-approved new users) only to logged-in readers? Jun 12 15:06:21 we have like 4 stages you need to get through, and they manage all of them, incl google-recaptcha Jun 12 15:06:38 "change UI" more radically, like "to register, please send an email to..." Jun 12 15:07:19 no thank you... we had a by hand registration for a while... that is becoming a dayjob Jun 12 15:07:41 also SPAM isn't attractive when noob posts are only visible to registered users and those registered users can 'approve' a new user by granting thanks Jun 12 15:08:09 90% of spam is only visible to moderators Jun 12 15:08:11 (dayjob) yeah I guess that's annoying and not sustainable Jun 12 15:08:55 we have akismet text validation that prevents a lots more as we like, but prevents a lot of spam too Jun 12 15:09:02 anyway, the big complaints are about blacklisting users from *reading* the forum Jun 12 15:09:47 well that is lifted by a mail to techstaff and won't change without a better idea Jun 12 15:09:59 that's completely unrelated to creating accounts, bypassing capchas etc Jun 12 15:12:08 hard to diff between just-readers and everything else Jun 12 15:12:16 gtg now, l8er Jun 12 15:12:20 o/ Jun 12 15:14:41 * DocScrutinizer05 ponders a set of URL-rewrite rules instead a redirect to 403 becoming effective for the blacklisted IPs only. So those URL-rewrites would catch all the "non read-only" pages only and block/redirect them Jun 12 15:16:29 ^^ Jun 12 15:16:37 this would allow browsing with Tor :) Jun 12 15:17:21 allowed URLs: http://talk.maemo.org/showthread.php?* http://talk.maemo.org/forumdisplay.php?f=* etc, all other URLs redirect to 403 still Jun 12 15:19:09 add on top the "quota" concept, probably most easily implemented by employing a carefully crafted rule for fail2ban Jun 12 15:20:14 so when a user exceeds 2 clicks/s for 10s, he gets banned for 15 minutes, or somesuch Jun 12 15:20:50 (adjust the citeria accordingly to match the real situation) Jun 12 15:21:14 (definitely needs some adjustment yeah ;)) Jun 12 15:21:21 e.g. "only 2 searches per minute" Jun 12 15:21:49 fail2ban rules are pretty flexible Jun 12 15:24:11 parazyd: that's the goal Jun 12 15:24:20 among others Jun 12 15:32:32 parazyd: why it is so silent on devuan-dev? Jun 12 15:36:36 i can't tell myself. it's like the email was simply ignored Jun 12 15:37:09 ignored, because i received replies to my other (unrelated) emails on that mailinglist Jun 12 15:37:38 weird Jun 12 15:37:44 :/ Jun 12 15:38:14 do I have our permission to forward that mail to maemo-devs? Jun 12 15:38:21 feel free Jun 12 15:38:27 ok, thanks Jun 12 15:38:36 also anyone that cares should reply Jun 12 15:42:28 which subject? Jun 12 15:42:46 https://lists.dyne.org/lurker/message/20170607.212507.1c785ac6.en.html Jun 12 15:43:06 uh, is maemo-devs still active? Jun 12 15:43:34 [devuan-dev] Building derivative distros on the Devuan Jenkins CI Jun 12 15:43:38 ? Jun 12 15:44:01 what? Jun 12 15:44:18 bencoh: why not? Jun 12 15:44:35 freemangordon: dunno, I'm just (positively) surprised :) Jun 12 15:45:19 well, will see when I send the mail :) Jun 12 15:45:48 I never registered because I didn't think it was tbh ... maybe I should then :) Jun 12 15:46:56 parazyd: I think teher's anly a few people on this globe who are competent to answer that quation in the email Jun 12 15:47:00 question* Jun 12 15:47:13 actually I can't remember if I have a registration either, or if it is with my abv.bg account Jun 12 15:47:36 DocScrutinizer05: those that are, seem to be ignoring the email, as they've replied to my other emails :/ Jun 12 15:54:03 bencoh: it looks like maemo-developers ML still works Jun 12 15:54:11 yay Jun 12 15:54:16 not that there is any activity recently :) Jun 12 15:54:47 DocScrutinizer05: are you subscribed to maemo-devs ML? Jun 12 15:56:20 Wizzup: while we are on it, I guess somebody should start releasing the packages, in terms of debian/changelog at least :) Jun 12 15:56:26 Yeah! Jun 12 15:56:48 I'm in hong kong right now. Intense work week behind me, and one more ahead of me. Home in 6 days. Jun 12 15:57:06 ok, no hurry Jun 12 15:57:18 It's unlikely I will pick it up before then, but that perhaps gives the devuan people time to our amprolla version up and running Jun 12 15:57:41 I'm happy to do the packaging stuff, but hope I don't have to work on devuan infra further (was fun, but also a distraction from the real goal) Jun 12 15:57:46 yeah, but I guess it is better we to be ready in advance, at least with couple of packages Jun 12 15:57:53 true. Jun 12 15:58:33 well, I guess if there is a need for some minor tweaks, parazyd will do them Jun 12 15:59:16 sure, and most of it should be in place. :) Jun 12 15:59:34 freemangordon: yep Jun 12 16:00:45 DocScrutinizer05: could you check if you receive a message on the ML from me? Jun 12 16:00:52 sure Jun 12 16:01:19 email anme? Jun 12 16:01:24 name Jun 12 16:01:31 "Fwd: Building derivative distros on the Devuan Jenkins CI" Jun 12 16:01:47 no, that one has only parazyd Jun 12 16:01:48 this has got eerie now - no network access at all on 3G Jun 12 16:02:04 sicelo: only on 3G? Jun 12 16:02:08 hmm, weird Jun 12 16:02:10 (2g working?) Jun 12 16:02:13 yea Jun 12 16:02:21 probably network issue then Jun 12 16:03:04 works on samsung. Jun 12 16:03:46 sicelo: it happened a couple of times to me, power cycle (with battery removal just in case) fixed it for me Jun 12 16:04:31 freemangordon: check https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/devuan-dev Jun 12 16:05:04 DocScrutinizer05: why? Jun 12 16:05:31 to check your account Jun 12 16:05:34 I am talking about maemo-developers ML, not devuan-dev :) Jun 12 16:05:42 oops sorry Jun 12 16:06:37 https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/devuan-dev Jun 12 16:06:53 ooh, still wrong Jun 12 16:07:01 no I'm not on developers ML Jun 12 16:07:46 ah, ok Jun 12 16:08:29 or I have no clue where the mails from that ML end in my dir-structure Jun 12 16:08:34 :) Jun 12 16:12:08 xes: ping Jun 12 16:13:58 freemangordon: pong Jun 12 16:16:54 xes: do you know if maemo-developers ML still works? and who is in charge of it? Jun 12 16:19:09 freemangordon: it's a long time since the last time i looked at it. I think that juiceme should take care of it Jun 12 16:20:16 xes: shall I ping him or you will do it? Jun 12 16:22:25 freemangordon: please ping him. At the moment i'm leaving a datacenter and i will be unavailable for a few hours Jun 12 16:22:33 ok Jun 12 16:22:38 juiceme: ping Jun 12 16:23:40 juiceme: Do you know if maemo-developers ML works? And if not, could you bring it back to life? Jun 12 16:44:12 sicelo: ((no network access at all on 3G)) check the BAND your provider supports. See if N900 does support it as well Jun 12 16:45:07 3G bands are a pita, 4G is a nightmare Jun 12 16:46:38 http://neo900.org/faq#networks for useful links Jun 12 16:47:05 đoes Jun 12 16:47:06 particularly https://en.wikipedia.org/wiki/UMTS_frequency_bands Jun 12 16:47:46 while usually frequency is sufficient, some are ambiguous so you want *band* Jun 12 16:49:25 it is supported. i don't know what's up really. Jun 12 16:49:49 i get 3G (unreliable) if i set N900 to Dual mode. 3G only gets no access at all Jun 12 16:50:20 take https://en.wikipedia.org/wiki/UMTS_frequency_bands#Deployments_by_region_.28UMTS-FDD.29 with a grain of slat! stuff changes faster than anybody can update wikipedia Jun 12 16:51:52 sometimes it helps to remove battery to make modem reboot hard Jun 12 16:52:34 wait, are you talking about data only or are calls also not possible? Jun 12 16:53:23 everything Jun 12 16:53:25 not unheard of that DHCP is fubar on one band only, and particular handover with keeping some DHCP info fails Jun 12 16:53:26 like i don't have coverage Jun 12 16:53:36 then it's not DHCP related Jun 12 16:54:18 sometimes it helps to swap the SIM Jun 12 16:54:37 brabd new sim Jun 12 16:54:43 hmm Jun 12 16:55:05 that might be the culprit though Jun 12 17:11:53 actually not. it's the N900 Jun 12 17:11:55 it's an old problem that I've never taken seriously because I would generally be home (where this network is) for only a few days at a time. Jun 12 17:11:55 now i'm here permanently and last two weeks i was using hotspot function off the s4. started using this new SIM today Jun 12 17:28:39 reflash modem with another locale version? **maybe** Jun 12 17:29:28 or it's actually a hw defect Jun 12 18:05:10 bug 6484 Jun 12 18:05:14 04Bug https://bugs.maemo.org/6484 GPRS/3G stops working but active connection is shown, manual reconnect needed Jun 12 18:51:55 yep, seen this a few times. That's why I suggested hard reboot Jun 12 18:52:24 it's seemingly related to 3G->2G handover Jun 12 18:52:37 and a modem problem Jun 12 18:52:49 try entering and leaving airplane mode Jun 12 18:52:59 might reboot the modem if you're lucky Jun 12 18:54:05 all that provides only temporary relief unfortunately Jun 12 18:54:05 it helps for sure ... but soon the 'problem' starts Jun 12 18:54:06 nothing unusual in syslog Jun 12 18:54:27 yeah, it's a modem problem Jun 12 18:54:50 fixing to 2G or 3G prolly helps Jun 12 18:55:03 i just wonder how the S4 (which is on same band as N900) can manage so well Jun 12 18:55:18 if airplane mode doesn't reboot the modem, a AT command via pnatd should do Jun 12 18:55:19 only to 2G Jun 12 18:55:25 fiing to 3G results in completely no reception Jun 12 18:55:44 S4 doesn't have that modem firmware bug Jun 12 18:56:04 yes Jun 12 18:56:19 so when your 3G is permanently broken then this bug doesn't match your problem Jun 12 18:57:07 this bug is about a rarely happening modem hangup when roaming to 2G from 3G in poor signal situation Jun 12 18:57:18 which gets fixed by a reboot Jun 12 20:21:26 * Maxdamantus usually does `killall csd` when connectivity seems to have been lost. Jun 12 20:22:05 * Maxdamantus might switch to 2G only for a while and see if he still has to do that. Jun 12 20:54:27 i will reflash the device during the course of the week to see what happens. i don't expect much, but i don't think i have much to lose Jun 12 21:11:38 fwiw, when my connection is lost, I see timeouts in the syslog a few minutes beforehand Jun 12 21:12:37 But I think those timeouts sometimes appear when the connection isn't lost, so it's probably that 3G connection dying before it tries to establish 2G as doc said. Jun 12 21:13:44 Jun 13 09:09:56 orcus daemon.notice cellular: csd[6666]: ISI_SMS .637686> ind_reg_status(): Net registration (ind) status:1 rc:0 Jun 12 21:13:47 Jun 13 09:09:56 orcus daemon.notice cellular: csd[6666]: ISI_SMS .643881> set_timeout(): Timeout 759 s event type:-1 Jun 12 21:24:19 yes that's always there Jun 12 22:47:25 you might try the non-global image. I have no particular clue which one image has a different modem firmware, except for the USA version **** ENDING LOGGING AT Tue Jun 13 03:00:00 2017