« first day (1159 days earlier)      last day (2958 days later) » 

tpu- by Mithrandir
Would !!/reboot work?
We've already tried.
AH, I see
13:08
tpu- by Floern
RIP Smokey until we resurrect you
!!/reboot
!!/tea
!!/blame
@Floern It's Floern's fault.
2
I'm sorry..
13:18
tpu- by Mithrandir
We're never going to set a new autoflagging record this way ...
Not today, no :/
But I got 9 more helpful flags from autoflagging before Smokey died...
well, we just need 11 more, it could still happen
averages would project something like 17 TP flags still during today
maybe it's almost an hour since I checked that
.oO(what I do in my copious spare time)
13:31
tpu- by Mithrandir
tpu- by Mithrandir
@ThomasWard you can't, it's down, scroll back
13:36
no smokey
@tripleee we can fix this, I believe... maybe
! !/dead bot
grabs Aurora
but I can tell you that luxmuscle\.com is in the URL blacklist
tpu- by Mithrandir
!!/location
@ThomasWard teward/Aurora
good you're alive for now
!!/alive
I hope.
13:39
@Mithrandir Of course
[ SmokeDetector | MS ] Manually reported answer: Avira AntiVir Personal Uninstall by Mark Berger on superuser.com
[ SmokeDetector | MS ] Manually reported question: T6 BLACK get-together by morinew sala on graphicdesign.stackexchange.com
@Undo @ArtOfCode emergency launch of Smokey on Aurora is confirmed, using new MS location that you gave me.
13:39
Restart: API quota is 9997.
since your RPi seems to have died, @Undo
"error during websocket handshake" though
@ThomasWard we are extremely happy, thanks!
ooo evil 302 response
blah at least Smokey is "up"
can't bind to MS websocket though, so you and I need to sync up, @Undo, at earliest convenience
(half-running mode)
What part doesn't work? The tpu- by xxxxx part?
@Glorfindel something about MS web sockets
13:41
@Glorfindel It didn't work because everyone who can already flagged it :P
bloody screen sessions >.<
Restart: API quota is 9990.
and we're back
13:45
!!/blame
@Glorfindel It's Thomas Ward's fault.
Lucky for him he provided a workaround :)
@ThomasWard yeah, but I was wondering how we would notice that.
@Glorfindel you probably wouldn't. It's being silently absorbed behind the scenes
though Undo and Art might be able to see
[ SmokeDetector | MS ] Blacklisted website in body, pattern-matching website in body, repeated URL at end of long post: malic damaging with the citrulline by Edgerephas on apple.stackexchange.com
@SmokeDetector k
13:48
Restart: API quota is 9959.
what died
!!/errorlogs 100
  File "ws.py", line 186, in <module>
    a = ws.recv()

  File "/usr/local/lib/python2.7/dist-packages/websocket/_core.py", line 293, in recv
    opcode, data = self.recv_data()

  File "/usr/local/lib/python2.7/dist-packages/websocket/_core.py", line 310, in recv_data
    opcode, frame = self.recv_data_frame(control_frame)

  File "/usr/local/lib/python2.7/dist-packages/websocket/_core.py", line 323, in recv_data_frame
    frame = self.recv_frame()

  File "/usr/local/lib/python2.7/dist-packages/websocket/_core.py", line 357, in recv_frame
oh
i need to error-capture that
[ SmokeDetector | MS ] Bad keyword in link text in body: What do you think about Yacht Charter Montenegro? by ahmet zengin on patents.stackexchange.com
I thought this was somehow fixed in the past, but maybe SO tweaked that again?
13:51
@tripleee there's been changes to WebSocket stuff in the code as well
tpu- by tripleee
so maybe something fubar'd.
gonna wrap a try/except in there somewhere
once my IDE loads
@ThomasWard I was running the sdml code base here to see if anything interesting would fly by while smokey was down, and I keep bumping into that same error
... beyond all repair?
I'd still rather it be captured and throw an error
than just outright crash
one of my things is to capture such errors :P
boo pycharm lag
13:53
@ThomasWard domain name should still work, you can keep config the same
[ SmokeDetector | MS ] Link at end of answer: Getting Web Designers to use Ubuntu by Pi Technologies Pvt. Ltd on askubuntu.com (@ThomasWard)
tpu- by tripleee
[ SmokeDetector | MS ] Link at end of answer, blacklisted user: How do I intall php7 on ubuntu? by Pi Technologies Pvt. Ltd on askubuntu.com (@ThomasWard)
tpu- by Mithrandir
@ArtOfCode i'm pointing it to the new metasmoke at metasmoke.charcoal-se, is that bad?
or is that OK?
bleh pycharm crashed
i'll deal with this later
at least Smokey is up for now
tpu- by Glorfindel
> Would not be caught as a question.
14:01
!!/blacklist-website privateequitycfo\.org
@Glorfindel Blacklisted privateequitycfo\.org
CI on 64566f1 succeeded. Message contains 'autopull', pulling...
Restart: API quota is 9852.
[ SmokeDetector | MS ] Manually reported answer: Avira AntiVir Personal Uninstall by Mark Berger on superuser.com
14:06
wasn't that already reported?
Reported in the room, but not to Smokey.
@Glorfindel ah of course
@SmokeDetector I guess this one was lost somehow
@Glorfindel Does a restart make it lose reports?
30 mins ago, by SmokeDetector
[ SmokeDetector | MS ] Manually reported answer: Avira AntiVir Personal Uninstall by Mark Berger on superuser.com
tpu- by Floern
Yeah, you see a Smokey restart message soon thereafter.
14:11
@Glorfindel Does a restart make it lose reports?
No, but I guess it crashed while processing the first report.
@ThomasWard I've only just changed the dns over on that, so it may not work
I don't know where it keeps track of already reported posts, but it seems the update of those records are fairly late in the process.
Not sure how long cloudflare ttl is
14:18
@Glorfindel if it didn't have a proper connection to metasmoke, that would have prevented the report from getting through
Yeah, point at metasmoke.erwaysoftware
So apparently when we changed ISP plans to a promotion to get the same service at a lot less money, part of that deal was "we're gonna disconnect you every night"
Thanks for taking over @ThomasWard, if you want you can just keep it for the weekend
@Undo this sounds like someone's gonna have a fun day shouting at ISPs
Yeah. We'll plan a tactical time to call, aka "when India is sleeping"
@tripleee probably needs updating to respond to heartbeats
:)
dunno, it seems to get that pretty regularly after fetching 100-something posts
14:25
@ArtOfCode Hard-coded DNS is fun. Especially when you run your own recursive resolvers.
But i'll point to ms.erway
I'm not even catching it properly yet, since I have to let it run for a while to see whether my code works
@ThomasWard but I don't :) It's on Cloudflare. I wouldn't even have a clue where to start going about running my own DNS for the domain
@ArtOfCode well I meant on my side
try something, 10 minutes, typo fix, 10 minutes, diagnose failure, 10 minutes
[ SmokeDetector | MS ] Bad keyword in title, pattern-matching website in body, repeated URL at end of long post: Frequensea Pro- Better fat loss by Potuza Cura on apple.stackexchange.com
14:26
@ArtOfCode CloudFlare instantly updates on their name servers
tpu- by tripleee
@ThomasWard does that include A records, though?
with my recursive resolvers in use by Aurora, it goes and fetches the IP data right from CloudFlare to the server
!!/test-q frequensea
> Would not be caught as a question.
14:27
@ArtOfCode every single non-cached lookup
!!/location
@ThomasWard Undo's Raspberry Pi
@Undo i think yours is back up
@ThomasWard not that I know what's cached and what isn't
but I'll stappit and taje it over
if you want
@ArtOfCode well here's how it works, if I run a query on my DNS nameserver, the cache is set to store records for 5 minutes and then forget. If it's cached, it's up to date within 5 minutes. Once it's gone from the cache, new data is fetched from the nameservers in use by the domain i'm looking for (in this case, CloudFlare for charcoal-se.org), and pulls whatever records I'm requesting
and returns them
only one hit so far but looks like it has potential as a brand name
14:28
so for Aurora, it's got 5-minute-cache on DNS lookup
then it goes and hunts for updated records
!!/stappit
Restart: API quota is 9690.
repointed
all is good.
@ArtOfCode Trust me, i've got updated records. Which actually match the IP that Undo gave me last night :)
which i know because I can check against my /etc/hosts where I temporarily hardcoded it :)
!!/location
@ThomasWard teward/Aurora
and we're all good :)
@ThomasWard Even for records with a TTL set longer than 300? So, say, if Cloudflare sends a TTL of 86400 with all its records, you'll still get the updates within 5 minutes?
14:30
@ArtOfCode well, it's a little more complex
basically, my recursive bind9 resolver receives a query for domain.tld.
it first checks its own cache, for the record
if it has the cached record, and it hasn't expired yet (I can force expire the cache every 5 minutes), it will return that cached lookup
if it has a cached record, and it has expired, it goes and recursively fetches updated record data, then stores that in the cache until that record expires
if it has no cached record at all, it goes and recurses out to the NameServers for the given domain, gets the record I'm looking for (say, A for domain.tld.), and adds it to the local cache
with a force-expire on the cached records every 5 minutes, it will always refresh the cache when the cache expires
same effect as restarting the bind9 process, but without having to respawn the entire set of processes
So. Some months ago, I set an A for metasmoke.charcoal-se.org to 54.xx.xx.xx. Let's say Cloudflare's TTL is 86400. Your server has that record cached (A metasmoke 54.xx.xx.xx ttl=86400). I update the A record to point at 37.xx.xx.xx instead. How long is it gonna take you to get that update?
@ArtOfCode since I instituted it yesterday, instantly. (It checks the SOA record of the domain first for its serial number, if it's newer it autorefreshes)
or rather, instantly upon the first lookup request.
and then refreshes its data every 5 minutes, provided a request is made to the bind9 server and the cache has expired
the other case is if I just rebooted the server, the cache is cleared, so it would instantly go and fetch updated records
@ThomasWard this is assuming you had the previous record, ttl 86400, cached?
@ArtOfCode assuming you changed the A record to 37.* yesterday, that's the first record it would ahve seen
assuming you set it this morning, the TTL is ignored because I have a different rule in my bind9 that expires the cached lookup data every 5 minutes
@ThomasWard Sure, but assuming, hypothetically, that you had the previous one cached - how long?
14:39
@ArtOfCode at most, 5 minutes.
because i clear and/or force-expire the cache on my bind9 every 5 minutes by force.
@ArtOfCode that said, part of the cache is to store the SOA record as well
so the first thing it does, cached or not, is check if the serial on the SOA has been incremented
not that I know what SOA does
[ SmokeDetector | MS ] Manually reported question: Frequensea Pro Up-to-date Be described by Potuza Cura on graphicdesign.stackexchange.com
tpu- by Glorfindel
and if the SOA serial has incremented, it pulls new data
14:40
!!/blacklist-keyword frequensea
@ArtOfCode SOA is the secret zone definition, for cache timeout, default TTLs, what is the nameserver that's ultimately authoritative, and a numerical serial number that indicates what 'revision' of the zone data we're at.
@tripleee Blacklisted frequensea
!!/test-q risingsupplements.com
> Pattern-matching website in body
----------
Body - Position 7-21: supplements.co
@ArtOfCode every time you change the data on CloudFlare, and it updates that DNS at CloudFlare's nameservers, the serial increments
so, at most, cache is 5 minutes if the serial hasn't changed
14:41
@ThomasWard hardly "secret" but it's like the handle to the rest of the information
CI on eb931bf succeeded. Message contains 'autopull', pulling...
if the serial has changed in the SOA record, which every zone has, then the cache is ignored anyways and new data fetched
@tripleee true, but you can't see the SOA on CloudFlare's interface
Restart: API quota is 9604.
only by direct queries
so from CloudFlare management interface perspective, SOA doesn't exist. But it really does behind the scenes.
@ThomasWard which are available to anyone with dig (or even, gasp, nslookup)
14:42
@ThomasWard Also, since you're the DNS expert around here... why is it that my domain registrar retains control of the NS records, but control of the other records is delegated to wherever those NS records point at?
pretty sure you've entered the dark world of deep DNS
@ArtOfCode registrar tells the domain TLD managers (i.e. ICANN for .net for example) what to point the NS to for a lookup for a given domain
@ArtOfCode the whole system is hierarchical, somebody higher up in the chain delegates to you the part which you need control over
take dark-net.io
you're not supposed to ask about that part, no one really knows how it works. It's like airplanes, 100% magic.
14:43
the actual recursion chain is, first: root domain servers (for the . zone). Second: TLD (io.), Third: my domain (dark-net.io.)
[ SmokeDetector | MS ] Offensive answer detected: Where did the term "Middle Ages" come from? by Daniel Keem on history.stackexchange.com
tpu- by Floern
the TLD's records system stores the NS data there and is populated by information made available from the registrars.
hence, the 'NS' records you set there show up in a whois on your domain, and DNS works
in a nut-shell
@ThomasWard whois is not authoritative, the root servers are
it's really complex, in reality. Do dig +trace A charcoal-se.org to see the whole chain
the vaguely magical part is the glue records
14:45
@tripleee yes, i know this, but the second part in that chain is not known until the registrars tell the TLD owners where to point the nameserver references
@tripleee glue can be ignored if you use a DNS server for another domain
the DNS server realizes you can't get there from here without some additional information, so it sends that along as well
@tripleee that part is true, but only insomuchas the nameserver records are wtihin the same domain
ns1.dark-net.net could be used in place of ns1.dark-net.io for example, and not need Glue at the dark-net.io level
mkay that's a long dig log
(though I do use glue recs)
@ArtOfCode yes, but if you specifically see the recursion trace you'll see all the individual hops taken, and the replies for each step along the way
Okay, next question: the following are the NS records for .:
.                       518400  IN      NS      B.ROOT-SERVERS.NET.
.                       518400  IN      NS      C.ROOT-SERVERS.NET.
.                       518400  IN      NS      D.ROOT-SERVERS.NET.
.                       518400  IN      NS      E.ROOT-SERVERS.NET.
.                       518400  IN      NS      F.ROOT-SERVERS.NET.
.                       518400  IN      NS      G.ROOT-SERVERS.NET.
.                       518400  IN      NS      H.ROOT-SERVERS.NET.
.                       518400  IN      NS      I.ROOT-SERVERS.NET.
14:48
@ArtOfCode godly overlord DNS master servers
that then point to the TLD specific zone data files along the first hop in the chain
@ThomasWard Sure. But, given that there are 13 of them, why do all root DNS requests from my LAN go to a.root-servers.net instead of distributing?
@ArtOfCode first available is used.
(mine goes to D. from my apartment, my own recursor on my laptop goes to D. E. A. G. randomly)
Looking at OpenDNS logs: RANK 1 DOMAIN a.root-servers.net REQUESTS 32744
the others don't even feature
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, pattern-matching website in body, repeated URL at end of long post, blacklisted user: Frequensea Pro Up-to-date Be described by Potuza Cura on graphicdesign.stackexchange.com
14:50
probably at large enough scales that it's not necessary to balance at the client level
@SmokeDetector k
@ArtOfCode doesn't matter which nameserver it goes to for root servers - they all have the same data pointers to other zones.
it's really not something to be concerned over.
(looking at OpenDNS logs also makes me realise that the vast majority of network traffic on this LAN is me, despite it being a house of 7 people)
@ThomasWard sure, I just wondered why it was :)
I get enough of DNS. The rest... Heathrow job.
I vaguely recall seeing some Windows-based configuration dialogs where you had to type in the root servers manually for your downstreams, so some folks who need to do that are probably just lazy and don't put more than they have to
back in a little while :P
14:54
@tripleee that's not what's happened here - all the routers are based on some variant of Linux
Huh. Actually, I wonder why my requests are showing up in the OpenDNS logs, when I'm not actually using OpenDNS... the rest of the LAN is, but my PC is set to use Google DNS
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, pattern-matching website in body, repeated URL at end of long post: T6 BLACK Retro eyeliner style by morinew sala on superuser.com
Anonymous
Wow. That's a lot of reasons why that post is spam.
@SmokeDetector k
@Mego the list can become so long that it doesn't fit in a chat message anymore.
The next message is usually from somebody here who wonders why the MS link isn't in the report...
Like this:
3 hours ago, by SmokeDetector
[ SmokeDetector ] URL in title, bad keyword in body, bad keyword in title, pattern-matching website in body, pattern-matching website in title, repeated URL at end of long post, blacklisted user: www.nutritionfit.org/ultra-fast-garcinia/ by Upperfic83 on graphicdesign.stackexchange.com
[ SmokeDetector | MS ] Link at end of answer: Can't hide video object by user3858966 on stackoverflow.com (@dorukayhan)
@SmokeDetector k
15:17
Is it just me, or is metasmoke now being incredibly slow?
Some parts of it, like the Search, yes.
Oh, the Dashboard now as well.
I'm getting 5+ seconds Rails-time on each request
Looks like MS is imitating the NS (the Dutch Railways).
Also, @Undo, Googlebot is crawling metasmoke
One of their nicknames is "Nooit Stipt" (never punctual).
15:22
@ArtOfCode what
@Undo Cloudflare analytics says 11 pages crawled in the last 24 hours
@ArtOfCode Happen to have timestamps / URLs?
Nope
Yandex is at it, too, though only 1 hit
which may have just been a hit on robots.txt
I suspect that's it hitting robots.txt, although I don't know why it'd do it 11 times a day
Oh, might just be scanning the root domain
i.e. the Charcoal website
that'd give it a few things to hit
15:32
@Art This might be a lot to ask, but (where) can I get feedbacks over what was auto-flagged in my name?
@Kyll Hm? Feedback on posts that your account cast a flag on? Should be shown on the per-user flag logs
click the link under flagging & you on the flagging dash
Indeed, thanks
Mh, 0 flags cast. I probably setup my preferences or conditions wrong.
Is it normal that I have nothing under "Sites" in preferences/new?
... Showed up after a couple refreshes. Blaming it on network goblins.
@Kyll known bug
[ SmokeDetector | MS ] Manually reported answer: Can I hear ultrasounds? by user7980 on health.stackexchange.com
those would be enterprising network goblins.
15:38
tpu- by Yvette
[ SmokeDetector | MS ] Few unique characters in body: what ever.what ever.what ever.what ever.what ever.what ever.what ever by NeoNaN on stackoverflow.com (@dorukayhan)
@SmokeDetector v
Yeah, metasmoke is cripplingly slow @Undo - 15970.6ms Rails time
approaching 20s TTG
I'm doing some stuff.
It gets all slow about when it starts swapping.
down temp
15:45
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, bad keyword in username, repeated URL at end of long post: skinsephora.com/borealis-cream-review/ ‎ by Borealis Cream on stackoverflow.com (@dorukayhan)
@SmokeDetector k
Did some thread tuning @ArtOfCode, should be better now.
Best thing might be to have MySQL and Passenger separate.
16:02
shrug you're the sysadmin :)
@SmokeDetector why
Title - Position 24-38: borealis-cream
Body - Position 243-257: borealis-cream, Position 293-307: Borealis Cream, Position 705-719: borealis-cream, Position 778-792: borealis-cream
Body - Position 211-811: <a href="http://skinsephora.com/borealis-cream-review/" rel="nofollow noreferrer">Borealis Cream</a> the explanations why our skin loses its physical property and why wrinkles and fine lines seem on our face. The primary reason for this can be aging and lots of different factors that races the method of aging and creating US look older than what we tend to are: sun exposure, free radical
@ArtOfCode @Undo last report's MetaSmoke link causes an error on MS.
27 mins ago, by Undo
down temp
other parts of MS do work at the moment.
!!/blacklist-website skinsephora\.com
16:11
@Glorfindel Blacklisted skinsephora\.com
@Glorfindel smokey never send the posts content to MS, hence its displays the development environment 404 error
huh?
Oh, right. We never sent it in this case.
@Glorfindel but they didn't when that post was reported and sent to MS, so it was never created on MS
CI on ed90c3d succeeded. Message contains 'autopull', pulling...
@ArtOfCode ah, makes sense.
16:12
Restart: API quota is 8934.
It's just that that page could show a normal error like 'Information of this post is not available in MetaSmoke'.
We could, yeah
[ SmokeDetector | MS ] Bad keyword with email in body: Found this out and he's the best! by Henrietta Hank Marion on drupal.stackexchange.com
Meh, I'll do that. Why not.
tpu- by Undo
16:14
'certified'
@SmokeDetector k
16:34
lurks the lurker
*lurkception results in division by zero, which in turn causes universal implosion*
[ SmokeDetector | MS ] URL in title, bad keyword in body, blacklisted website in body, blacklisted website in title, repeated URL at end of long post: cluesforhealth.com/Derma-Reflexion/ by SherryLop on graphicdesign.stackexchange.com
tpu- by Glorfindel
[ SmokeDetector | MS ] Few unique characters in body, no whitespace in body, repeating characters in body: Will deleting a product affect our google points? by Rebecca on wordpress.stackexchange.com
@SmokeDetector v
[ SmokeDetector | MS ] Few unique characters in answer, repeating characters in answer: Safety of using a luggage with the prop. 65 warning by C.h.o.b on health.stackexchange.com
tpu- by Glorfindel
16:59
@Glorfindel Done
Excellent!
17:17
@Undo do we have code to alert when Smokey is down for a substantial amount of time, sort of like a fire alarm bell that goes out to people who can rapid-deploy an instance?
I coulda sworn we had such a system
metasmoke will send an email
what timeframe does it consider substantial
five minutes or so
because supposedly Smokey was down for two hours and no email went out, that I could see
and I think i'm on that notice list?
@ThomasWard Only if yours is the one that goes down.
We can add you to the cc list if you want, github.com/Charcoal-SE/metasmoke/blob/master/app/mailers/…
Last night there wasn't an email because of metasmoke issues. Perfect storm.
17:19
ah, that explains it
@Undo if you could add me to the CC list that'd be great
@ThomasWard Email?
because if the Smokey manager isn't online, we'd need to have a temporary other instance up
thanks
17:22
@Undo thanks
We should probably have a way to take those addresses out of code, should be on user models.
indeed

« first day (1159 days earlier)      last day (2958 days later) »