Back to Home / #linode / 2019 / 06 / Prev Day | Next Day
#linode IRC Logs for 2019-06-11

---Logopened Tue Jun 11 00:00:19 2019
00:11-!-Axle22 [~oftc-webi@97.103.184.30] has joined #linode
00:11-!-Axle22 is "OFTC WebIRC Client" on #linode
00:12<Axle22>It says your plans come with 40GB of data transfer *IN*. What is the pricing once that limit is exceeded?
00:16<rsdehart>Axle22: https://www.linode.com/docs/platform/billing-and-support/network-transfer-quota/#how-overages-work
00:18-!-kaare__ [~kaare@ip-5-186-247-169.dhcp.fibianet.dk] has joined #linode
00:18-!-kaare__ is "Kaare Rasmussen" on #linode
00:25-!-ELA_ [~oftc-webi@122.165.140.47] has joined #linode
00:25-!-ELA_ is "OFTC WebIRC Client" on #linode
00:25<ELA_>Good Morning!!
00:25<ELA_>In Linode do you provide any windows service
00:29<rsdehart>!winode
00:29<linbot>It is possible to run Windows on !kvm Linodes. Here's a set of unofficial instructions: https://github.com/linode/docs/pull/501#issuecomment-232414947
00:30<rsdehart>it's not officially supported
00:30<rsdehart>ELA_: ^^
00:30<ELA_>Thank you
00:32-!-fil_ [~oftc-webi@116.226.186.165] has quit [Remote host closed the connection]
00:45-!-kaare_ [~kaare@ip-5-186-247-169.dhcp.fibianet.dk] has joined #linode
00:45-!-kaare_ is "Kaare Rasmussen" on #linode
00:45-!-kaare__ [~kaare@ip-5-186-247-169.dhcp.fibianet.dk] has quit [Read error: Connection reset by peer]
00:48-!-Axle22 [~oftc-webi@97.103.184.30] has quit [Quit: Page closed]
01:07-!-ELA_ [~oftc-webi@122.165.140.47] has quit [Quit: Page closed]
01:12<linbot>New news from community: I am having a problem with updating Wordpress & Plugins <https://www.linode.com/community/questions/18326>
01:22<linbot>New news from community: Problems migrating vultr to Linode <https://www.linode.com/community/questions/18327>
02:05-!-Bonz [~oftc-webi@124.107.173.50] has joined #linode
02:05-!-Bonz is "OFTC WebIRC Client" on #linode
02:05<Bonz>Hi
02:05<Bonz>Anyone from technical support?
02:10-!-Bonz [~oftc-webi@124.107.173.50] has quit [Quit: Page closed]
02:12<Zr40>!ops
02:12<linbot>Users with ops are employees of Linode, and know what they're talking about. The rest of us are the ever-so-helpful(?) community. Official Linode contact information: https://www.linode.com/contact
02:13<dwfreed>they're already gone
02:23-!-joecool|mobile [~joecool@c-174-57-44-238.hsd1.nj.comcast.net] has joined #linode
02:23-!-joecool|mobile is "Joe" on #ck #linode
02:29-!-nate [NBishop@d-207-255-41-254.paw.cpe.atlanticbb.net] has joined #linode
02:29-!-nate is "Nathan" on #linode #php
04:10-!-blaboon [~blaboon@00026ecf.user.oftc.net] has quit [Quit: Bye o/]
04:11-!-blaboon [~blaboon@00026ecf.user.oftc.net] has joined #linode
04:11-!-blaboon is "Bradley LaBoon" on #linode
04:11-!-mode/#linode [+o blaboon] by ChanServ
05:47-!-Luckst0r [~luckst0r@luckst0r.soho.on.net] has joined #linode
05:47-!-Luckst0r is "luckst0r" on #linode
06:51-!-jaskal [jaskal@jaskal.user.oftc.net] has left #linode []
06:53-!-darwin [d@melik.windwireless.net] has quit [Ping timeout: 480 seconds]
06:58-!-anomie [~anomie@00018802.user.oftc.net] has joined #linode
06:58-!-anomie is "Anomie" on #linode
07:46-!-spiki [~spiki@0001014f.user.oftc.net] has joined #linode
07:46-!-spiki is "Nenad Spirkoski" on #linode
07:53-!-The-spiki [~spiki@0001014f.user.oftc.net] has quit [Ping timeout: 480 seconds]
08:34<squidly>Are there any linode employees on that can help me with a DNS issue?
08:35<nyancat>i hope i'm doing this right
08:35<nyancat>!ops squidly
08:35<linbot>Users with ops are employees of Linode, and know what they're talking about. The rest of us are the ever-so-helpful(?) community. Official Linode contact information: https://www.linode.com/contact
08:35<grawity>they tend to appear more often after the problem has been explained
08:35<grawity>described*
08:35<nyancat>squidly: you can always just ask your question and someone will try to help though
08:35<Zr40>!ask
08:35<linbot>If you have a question, feel free to just ask it -- someone's always willing to help. If you don't get a response right away, be patient! You may want to read http://alexfornuto.com/how-to-ask-for-help-on-irc/
08:35<grawity>though in general, there are two kinds of problems: those you can file a support ticket for, and those you don't need linode employees for
08:36<squidly>It's a simple issues. I Linode is a slave on for my DNZ sones. I've updated my master zone (with a difference zone Serial number) and the Linode DNS servers are not taking the update.
08:37<squidly>Ive had a ticket open for a day with no responce
08:39-!-squidly [~squidly@2600:3c03::f03c:91ff:fead:ed05] has quit [Quit: Lost terminal]
08:40-!-squidly [~squidly@2600:3c03::f03c:91ff:fead:ed05] has joined #linode
08:40-!-squidly is "squidly" on #linode
08:40<squidly>oops wrong screen to close
08:41<Woet>squidly: did you whitelist the Linode IPs for AXFR?
08:44<grawity>I was going to say, have you done the obvious like checking your DNS master's logs for zone transfer attempts
08:45<grawity>but got distracted by IRL work
08:47<Peng>Did you send notifies? What's the refresh interval?
08:47<squidly>Yep. It took one of the updates
08:48-!-eyepulp [~eyepulp@c-71-239-4-144.hsd1.il.comcast.net] has joined #linode
08:48-!-eyepulp is "eyepulp" on #linode
08:48<Peng>"one of the updates"?
08:48<squidly>with a new SN on the zone file It should grab the new file.
08:49<squidly>I was doing a couple of changes as I was moving my servers around. The and every zone change is given with a new Serial number. Using the name servers from linode should show the new SN when I do a SOA querry
08:49<Peng>...
08:50<squidly>??
08:51<Peng>Have cat can't type. Please answer questions
08:52<squidly>dig @ns1.linode.com -t SOA myzone shows the old SN. That is telling me that the Linode servers are not refreshing the zone.
08:52<grawity>1) what is your SOA "refresh" period? 2) have you moved the addresses of the master server itself?
08:52<squidly>Yes they are whitelisted and permitted to querry. I have seen the attempts in my log files for the axfers
08:52<grawity>do those xfers succeed?
08:53<squidly>the master IP has not been changed
08:54<squidly>refres was 28800. Regardless the slave should transfer the zone file when the SN is updated
08:55<Peng>By the way, nsN.linode.com do caching, but IIRC it's not for very long.
08:55<grawity>it should if the master sends DNS NOTIFY to replicas (which I've seen breaking occassionally)
08:56<squidly>That is why I was looking at the Linode nameservers. If they dont have the most recent file, they cant proved that to anyone after a direct querry.
08:56<squidly>grawity: That is what I'm thinking. I think the DNS servers for linode are stuck for my zone for some reason.
09:08<Peng>What if the new version of the zone was invalid?
09:08<Peng>Not extremely likely, but possible.
09:10<grawity>if the server loaded it and even logged a zone transfer
09:10<squidly>I checked my zone files with named-checkzone before I do that.. I had an isssue years ago where I was on the other side
09:10<grawity>though I wonder whether it actually logged a zone transfer *to Linode*
09:11<grawity>are those Linode servers queryable directly?
09:11<squidly>yes
09:14<Peng>What if *Linode* thinks the new zone is invalid? Though it sounds unlikely.
09:16<squidly>Peng: that's what I am thinking is going on
09:16<squidly>I've had that issue in the past, both with Linode and other providers.
09:18<grawity>hmm, maybe if you use some very new rrtypes and Linode uses some very old nameservers
09:19<Peng>BIND's supported RFC 3597 forever though. So that shouldn't happen easily.
09:20<squidly>grawity: my internal DNS servers (UBUNTU LTS) has some oddities where the slaves journal became corupt
09:46<nate>squidly: manually edit the zone files any time recently?
09:48<squidly>nate: yes.
09:53-!-mortie [~oftc-webi@ip72-221-28-225.ph.ph.cox.net] has joined #linode
09:53-!-mortie is "OFTC WebIRC Client" on #linode
09:53-!-mortie [~oftc-webi@ip72-221-28-225.ph.ph.cox.net] has quit []
10:03<nate>squidly: That would be likely why, not an ubuntu thing so much as bind, started a while back with one of the Bind9 updates, manual updates corrupt the journals, have to delete the .jnl files and restart bind
10:03<nate>I had a couple domains do it after I manually edited an update key into them for acme.sh updates (cause webmin was breaking it when I tried specifying it in there)
10:08<squidly>nate: yea that is what I had to do on my internal DNS zones
10:59<@mcintosh>!wx kphl
10:59<linbot>mcintosh: [metar] OBS at KPHL: 68.0F/20C, visibility 10 miles, wind 16.11 mph, chill 67.58F (altimeter: 29.98) [KPHL 111454Z 33014KT 10SM FEW034 SCT250 20/11 A2998 RMK AO2 SLP153 T02000111 ]
11:02<dwfreed>!wx KAMN
11:02<linbot>dwfreed: [metar] OBS at KAMN: 69.8F/21C, visibility 10 miles, wind 4.60 mph, chill 71.57F (altimeter: 30.19) [KAMN 111459Z AUTO 26004KT 10SM CLR 21/08 A3019 RMK AO2 T02060084]
11:02<dwfreed>I win?
11:27<LouWestin>I just read that Firefox is looking to offer a VPN and cloud storage subscription service sometime in the fall.
11:27<LouWestin>https://www.theverge.com/2019/6/10/18660344/firefox-subscription-paid-service-vpn-cloud-storage-release-date
11:45<cruxeternus>Yea, they're moving to the Freemium model
11:45<cruxeternus>which is one step away from Lootboxes
11:46<millisa>looks like MDS maintenance tickets are going out.
11:46<LouWestin>For me, I wouldn’t use it. I like keeping things separate.
11:47<LouWestin>For the intel thing?
11:47<millisa>yeah
11:47<LouWestin>Ok
11:47<Peng>If you mutate a Linode, will it land on a host that doesn't need MDS maintenance?
11:47<cruxeternus>I got mine several days ago
11:47<millisa>I think all of them I've seen start as early at 10p edt and go till 4
11:47<millisa>they don't say. sec and i'll post a screen of their notice
11:48<millisa>https://vomitb.in/sMVyvCzgDt
11:49<millisa>looks like my first account's notice was about 9 hours ago, most recent was a little over an hour, so I'm guessing they are still generating
11:50<LouWestin>I haven’t gotten an email yet, but that should only affect one of mine
11:51<millisa>so far its just been one ticket per account. some accounts appear to have some that aren't impacted.
11:52<cruxeternus>presumably the AMD ones?
11:53<millisa>i have some intels that aren't recent spinups or migrations that don't have a notice on them
11:58<millisa>well, I guess this isn't a great test. tried resizing one of my test 1gb nodes to 2gb. ended up on an amd box. the maintenance notice did go away in the manager.
12:00<cruxeternus>"Clever girl"
12:03<millisa>going back I still was an on amd. wouldn't surprise me if they have a preference list for host it tries to pick first.
12:04-!-CodeMouse92 [~JasonMc92@00025241.user.oftc.net] has joined #linode
12:04-!-CodeMouse92 is "Jason C. McDonald" on #packaging #linode #c++
12:16<Peng>Mm. If so, I wonder what the percentage of "patched Intel" vs "AMD" is.
12:47-!-Shentino [~shentino@50.35.136.147] has quit [Remote host closed the connection]
12:50-!-Shentino [~shentino@50.35.136.147] has joined #linode
12:50-!-Shentino is "realname" on #mm #linode #tux3
12:51*dwfreed remembers mega reboot 2013
12:51<millisa>text reads kinda like that one
12:51<dwfreed>good old XSA-7, -8, and -9
12:51<dwfreed>now the numbers are like 250
12:53<LouWestin>Has there been any AMD patches?
12:53<dwfreed>AMD is not vulnerable to MDS and Zombieload
12:53<LouWestin>I don’t recall any since I’ve been here
12:53<dwfreed>AMD was vulnerable to Meltdown and Spectre, and I believe they also issued microcode mitigations
12:54<LouWestin>Ok I remember XEN patches but that’s hyper visor
12:54<LouWestin>Oh yeah
12:54<Peng>dwfreed: And now Linode runs KVM, which is part of Linux and therefore doesn't properly announce or organize vulnerabilities, so they're uncountable. \o/
12:54<LouWestin>I think it was certain AMD processors
12:55<LouWestin>Would kvm be patched in the kernel patches?
12:55<dwfreed>yes
12:56<LouWestin>Ok I’m on the Deb mailing list so I see a kernel patch once in awhile
12:58<chesty>I take MDS and Zombieload can't be hot patched or whatever the term is?
12:59<dwfreed>I believe there are new microcodes
12:59<dwfreed>while you can do it, it's generally not a good idea to load microcode once you've made it to userspace
13:00<Peng>chesty: So far people are getting emails about maintenance for *some* of their Intel Linodes.
13:01<Peng>I suppose they're handling some of it with live migrations?
13:01<dwfreed>I've gotten emails for all OFTC Linodes, as well as my personal Linode
13:01<Peng>D:
13:01<chesty>i have one of those emails.
13:01<dwfreed>Peng: I don't think so; the behavior change at the target host would be too risky, and the guest kernel would not know about it
13:02<Peng>dwfreed: That makes more sense, but raises the question of why not every Intel Linode is getting rebooted yet.
13:02<Peng>Maybe they're doing it in waves.
13:03<dwfreed>most likely waves, just because there's so many hosts
13:03<Peng>Yeah. :X
13:03<dwfreed>remember, mega reboot 2013 took like 2 weeks to get every host
13:04<dwfreed>as Linode exposes the full host CPU to the guest, the only completely safe live migration is to another host with the same processor and microcode
13:04<Peng>https://blog.digitalocean.com/may-2019-intel-vulnerability/ DigitalOcean seems to have done it without rebooting everyone
13:04<nyancat>So I can't seem to update the rDNS on my ipv6.
13:04<Peng>nyancat: What's happening?
13:04<nyancat>nothing, actually.
13:04<Peng>Specifically, DigitalOcean claims to be done, and I didn't get rebooted. :P
13:04<dwfreed>nyancat: what name are you trying to use, and what's the expecting IPv6 address that should be at that name?
13:05<Peng>nyancat: You click the button and nothing happens?
13:05<dwfreed>s/expecting/expected/
13:05<nyancat>Yeah, I typed the record i had, clicked save, and absolutely nothing happens
13:05<nyancat>buuuuuuuuuuuuuut it just
13:05<nyancat>well
13:05<nyancat>never mind.
13:05<dwfreed>took a minute to load?
13:05<nyancat>a hot minute, yeah
13:05<nyancat>more like 3 or 4 minutes
13:05<dwfreed>Peng: does DO still hide the host CPU?
13:06<nyancat>i dont think so
13:06<Peng>dwfreed: Define hide. /proc/cpuinfo looks about as revealing as Linode.
13:06<dwfreed>Peng: so it says actual CPU model number, instead of "QEMU Virtual CPU" ?
13:06<nyancat>model name : Intel(R) Xeon(R) CPU E5-2650 v4 @ 2.20GHz
13:07<nyancat>Taken from a droplet in Toronto
13:07<millisa>If it's of interest - this is an account that has a mix of notices/no-notice and has some intels that weren't part of that first wave ticket
13:07<millisa>https://ibin.co/4kFrNRSSPND0.png
13:07<nyancat>yeah i've yet to get a single notice
13:07<nyancat>mind you, i only have two linodes
13:07<Peng>nyancat: Lucky!
13:07<Peng>model name : Intel(R) Xeon(R) CPU E5-2650L v3 @ 1.80GHz
13:07<dwfreed>lol L model
13:07<Peng>ikr
13:08<dwfreed>the L models *sucked*
13:08<nyancat>lol what
13:08<@mcintosh>not all of the notices have been sent
13:08<millisa>we figured
13:08<nyancat>hey
13:08<nyancat>hey guys
13:08<nyancat>let's use
13:08<millisa>!point mcintosh
13:08<linbot>millisa: Point given to mcintosh. (15) (Biggest fan: csnxs, total: 6)
13:08<nyancat>hey guys, listen
13:08<nyancat>let's use L cpus
13:08<@mcintosh>millisa: out of curiosity, do those other linodes have more specific maintenance tickets?
13:08<nyancat>in a hosting...
13:08<nyancat>you know what, no.
13:08<Peng>Honestly, I've had good and bad experiences with contention and CPU performance on DO with those CPUs.
13:08<millisa>1 had a recent migration and is an amd. other two don't have any notices and aren't interesting really.
13:09<dwfreed>Peng: you probably got some empty hosts
13:09<nyancat>i have one of these in mine. model name : Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
13:09<dwfreed>nice
13:09<nyancat>mmm. sweet gold.
13:10<Peng>dwfreed: Well, I'm not gonna complain about that. :D
13:10<dwfreed>nyancat: 20 cores + HT, and 4 way scalability
13:10<dwfreed>host could have up to 3 TB of RAM
13:11<nyancat>in my mail server... model name : Intel(R) Xeon(R) CPU E5-2697 v4 @ 2.30GHz
13:12<dwfreed>18 cores + HT, 2 way scalability
13:13<dwfreed>(all E5 processors are max 2 socket)
13:14<Peng>Which is probably Intel'x excuse for charging like twice as much
13:14<dwfreed>the Xeon Platinum 8280 is 28 cores + HT, and 8 way scalability
13:15<dwfreed>you could have a host with 448 threads in it
13:15<dwfreed>for the low low price of 100k after all costs are taken into consideration
13:15<dwfreed>or more
13:15<dwfreed>(it's 80k just for the processors)
13:16<Peng>"TDP: 205 W"
13:16*LouWestin whistles...
13:16-!-fstd [~fstd@xdsl-78-35-64-33.nc.de] has joined #linode
13:16-!-fstd is "fstd" on #oftc #linode #debian #kernelnewbies
13:16<nyancat>Huh. Guess it takes a while to update DNS now.
13:17<dwfreed>always has
13:17<dwfreed>!dns15
13:17<linbot>The DNS Manager reloads your zones every quarter hour. If you just added/changed a record you'll need to wait until the next refresh before the changes will be reflected by Linode's nameservers.
13:17<nyancat>I've never had to wait. *fusses*
13:17<dwfreed>that same quarter hour interval applies to the rDNS zones as well
13:17<Peng>Plus up to 15 minutes of caching if you keep running "dig"
13:18<nyancat>I forgot to change the rDNS on my IPs when I created this server so my mail keeps going to spam
13:18<nyancat>google is looking for the host of the mail server and the ptr record to match and they don't whoops.
13:18<dwfreed>Peng: oh yeah, 1640 watts of heat, just for your processors
13:19<dwfreed>make a stove heatsink for the processors, cook some stuff while it runs
13:21<millisa>step 1 - put brownie mix in mug. step 2 - open server and put mug in void near processors. step 3 - wait for dusty brownies, forget which rack you were in. it's fine, server room smells all chocolatey.
13:23<nyancat>nah. pizza.
13:24-!-fstd_ [~fstd@xdsl-87-79-163-152.nc.de] has quit [Ping timeout: 480 seconds]
13:35<LouWestin>Or incorporate them into the hvac system
13:35<millisa>and screw up the jerky racks!?
13:38<LouWestin>Ha ha, set aside for the jerky racks
13:38<LouWestin>Some for pizza, etc.
14:03-!-V-Pariah [viciouspar@c-24-61-207-78.hsd1.ma.comcast.net] has quit [Read error: Connection reset by peer]
14:20<nyancat>Hasn’t the server pizza been done already?
14:23-!-V-Pariah [viciouspar@c-24-61-207-78.hsd1.ma.comcast.net] has joined #linode
14:23-!-V-Pariah is "Vicious Pariah" on #linode
14:33-!-alan [~oftc-webi@107-179-235-11.cpe.teksavvy.com] has joined #linode
14:33-!-alan is "OFTC WebIRC Client" on #linode
14:40-!-alan [~oftc-webi@107-179-235-11.cpe.teksavvy.com] has quit [Quit: Page closed]
15:12<v0lksman>is there a way to force the migration without having to rebuild my node?
15:12<akerl>Resize for a second and then resize back?
15:13-!-Strykar [~wakka@strykar.user.oftc.net] has quit [Quit: /quit]
15:14-!-Strykar [~wakka@strykar.user.oftc.net] has joined #linode
15:14-!-Strykar is "vector" on #linode #bitrig #bitlbee
15:15<smccabe>Cloning and swapping the IP would work too.
15:16<akerl>You’d lose the v6 address that way, right?
15:17<dwfreed>yeah
15:17<smccabe>Yes you would, but on the upside you wouldn't have the downtime from migration resizing. And Network helper should do its think and get the new v6 setup correctly.
15:17<dwfreed>that part hasn't changed since we left
15:18<dwfreed>smccabe: a live clone often isn't a good idea
15:18<akerl>Will network helper update DNS?
15:18<dwfreed>because you have no disk consistency
15:18<smccabe>dwfreed True, bould would be faster.
15:18<smccabe>*but would*
15:19<dwfreed>it's not really faster if the data is corrupt when it gets there
15:20<smccabe>dwfreed I ment if it was powered off when cloned.
15:20<dwfreed>a powered off clone is no different than a migration
15:21<smccabe>should have included that. Depends. I have personally noticed faster speeds when cloned from powered off vs migration ¯\_(ツ)_/¯
15:21<akerl>Sure, but there’s not a “start the migration now” button that’s self-service. You have to resize, so it’s double-hop
15:21<dwfreed>smccabe: that is simply host variation
15:21<smccabe>also what akerl mentioned
15:21<smccabe>Double hop vs one way
15:22<dwfreed>*shrug*
15:23-!-JoeK [~Joseph@ny1.us.makaiwell.com] has quit [Ping timeout: 480 seconds]
15:23<dwfreed>but a double resize would probably be faster than waiting for DNS to update if you care about IPv6 and are using SLAAC addresses for that
15:32-!-alan [~oftc-webi@107-179-235-11.cpe.teksavvy.com] has joined #linode
15:32-!-alan is "OFTC WebIRC Client" on #linode
15:33<alan>https://bpaste.net/show/60a5522b5a5e
15:33-!-mmustac [~matt@0001b77f.user.oftc.net] has joined #linode
15:33-!-mmustac is "matt" on #linode
15:33<alan>this mysql query isn't working for the life of me, it's the order and or paranthesis missing or mispalced
15:33<alan>misplaced **
15:33<alan>it doesn't target only the orders with user_id = 2
15:33<alan>it targets all
15:34<alan>I think perhaps an IN clause would fix the issue but I'm not an expert with sql queries
15:36<millisa>WHERE np_affiliate.ref_user_id = 2 is not WHERE np_affiliate.user_id = 2 ?
15:38<alan>no
15:38<alan>i currently have only one other order that hgas revoked set to 1
15:38<alan>user_id = 1 for that row
15:38<alan>the other one is user_2 but my querie targets both orders idk why?
15:38<millisa>I mean your where clause does not have user_id=2
15:39<alan>sorry ref_user_id
15:44<alan>i just dont get it tbh
15:55-!-JoeK [~Joseph@ny1.us.makaiwell.com] has joined #linode
15:55-!-JoeK is "Joseph A Kankowski" on #oftc #linode
16:03-!-duckydanny [~duckydann@2600:3c02::f03c:91ff:fec7:36d9] has quit [Quit: ZNC - http://znc.in]
16:06-!-duckydanny [~duckydann@2600:3c02::f03c:91ff:fec7:36d9] has joined #linode
16:06-!-duckydanny is "Dan" on #tor-project #moocows #linode #debian
16:06<dwfreed>alan: strings need to be single quoted
16:07<dwfreed>also you can remove that paren by using np_orders.status IN ('paid', 'subscribed')
16:09<dwfreed>granted, I did this test in sqlite, but this works:
16:10<dwfreed>SELECT SUM(np_orders.paid * 0.44) AS actual FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_affiliate.ref_user_id = 2 AND np_orders.revoked = 0 AND np_orders.status IN ('paid', 'subscribed');
16:10<dwfreed>and this doesn't:
16:10<dwfreed>SELECT SUM(np_orders.paid * 0.44) AS actual FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_affiliate.ref_user_id = 2 AND np_orders.revoked = 0 AND (np_orders.status = "subscribed" OR np_orders.status = "paid");
16:10<dwfreed>(also the second form but with single quotes worked)
16:14<dwfreed>in MySQL "JOIN" without specifying a type is automatically an INNER JOIN
16:23<alan>OK fixed my other issue lets see if i can fix this last one with you all
16:23<alan>sorry went MIA
16:23<alan>SELECT SUM(np_orders.paid * 0.44) AS actual FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_affiliate.ref_user_id = 2 AND np_orders.revoked = 0 AND np_orders.status IN ('paid', 'subscribed');
16:24<alan>this gives me a closer number to my figure desired actually!
16:24<alan>nvm i lied
16:25<alan>still not working
16:27<dwfreed>try this: SELECT SUM(paid * 0.44) AS actual FROM np_orders WHERE np_orders.user_id IN (SELECT user_id FROM np_affiliate WHERE ref_user_id = 2) AND revoked = 0 AND status IN ('paid', 'subscribed');
16:27<dwfreed>do you get the same answer with this as the select you just did?
16:28-!-kaare_ [~kaare@ip-5-186-247-169.dhcp.fibianet.dk] has quit [Read error: Connection reset by peer]
16:28<alan>100% same answer
16:28-!-kaare_ [~kaare@ip-5-186-247-169.dhcp.fibianet.dk] has joined #linode
16:28-!-kaare_ is "Kaare Rasmussen" on #linode
16:28<alan>it still targets both ref_user_id 1 & 2 :/
16:29<dwfreed>then your mysql server is broken
16:29<dwfreed>because there's no way that second query would return a user from np_affiliate whose ref_user_id is 1
16:30<dwfreed>try just the subquery: SELECT user_id, ref_user_id FROM np_affiliate WHERE ref_user_id = 2;
16:32<dwfreed>and maybe pastebin the result of that ^
16:34<alan>https://bpaste.net/show/c376e8a3dfcc
16:34<alan>there
16:35<dwfreed>are those all the user_ids you expected?
16:36<dwfreed>are there any you didn't expect, or are there some missing?
16:36<alan>those are fine
16:37<dwfreed>also now that there's that list, I can remove the subquery; SELECT SUM(paid * 0.44) AS actual FROM np_orders WHERE user_id IN (4, 5, 12, 13, 14, 18, 19, 21) AND revoked = 0 AND status IN ('paid', 'subscribed');
16:38<alan>same result 192.50
16:38<alan>should be 196.50
16:38<dwfreed>so I think your query is fine, and your expectation is broken
16:38<alan>no
16:38<alan>look
16:39<alan>my expectation is fine
16:40<dwfreed>SELECT np_orders.paid, np_orders.user_id, np_affiliate.ref_user_id, np_orders.revoked, np_orders.status FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_affiliate.ref_user_id = 2 AND np_orders.revoked = 0 AND np_orders.status IN ('paid', 'subscribed');
16:41<dwfreed>if your dataset is small enough, leave off the where clause and pastebin the whole thing
16:44<dwfreed>alan: ^
16:45<alan>sorry I need a little break I'm going crazy
16:45<alan>spent too much time on this
16:52<alan>it still doesnt work
16:52-!-anomie [~anomie@00018802.user.oftc.net] has quit [Remote host closed the connection]
16:52<dwfreed>pastebin the result of the last select I gave, with or without the where clause?
16:53<alan>SELECT SUM(paid * 0.44) AS actual FROM np_orders WHERE np_orders.user_id IN (SELECT user_id FROM np_affiliate WHERE ref_user_id = 2) AND revoked = 0 AND status IN ('paid', 'subscribed');
16:53<alan>doesnt give me the right answer even if i put all revoked to 0 in phpmyadmin
16:54<dwfreed>SELECT np_orders.paid, np_orders.user_id, np_affiliate.ref_user_id, np_orders.revoked, np_orders.status FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_orders.paid = 9.09;
16:54<dwfreed>^ give me the result of that one too
16:55<alan>gives 0 results ^
16:55<dwfreed>do you have any records where paid is less than 9.09 that you'd expect to be included?
16:56<alan>2 yes
16:56<alan>4.99
16:56<dwfreed>both 4.99?
16:56<alan>2 that are 4.99
16:56<dwfreed>then the answer can't be off by exactly 4 dollars
16:57<dwfreed>because 9.09 is what paid would have to be to produce 4 in the formula for summing
16:57<dwfreed>and if you have only 2 orders less than 9.09, and they're both 4.99, they sum to more than 9.09 before doing the * 0.44
16:57<alan>you query guivesm e 204.81
16:57<alan>me **
16:57<alan>when it should be 208.33
16:58<dwfreed>you said it was 192.50 above
16:58<dwfreed>and that it should be 196.50
16:58<alan>i added records since
16:58<alan>as i stopped to work on something else
16:58<dwfreed>pastebin 'SELECT np_orders.paid, np_orders.user_id, np_affiliate.ref_user_id, np_orders.revoked, np_orders.status FROM np_affiliate INNER JOIN np_orders ON np_orders.user_id = np_affiliate.user_id'
16:58<alan>diff of 3.,52
16:58<dwfreed>(the result of it, obviously)
16:59<alan>which means theres 1 package
16:59<alan>at 7.99 that is messing something up
16:59-!-darwin [d@melik.windwireless.net] has joined #linode
16:59-!-darwin is "Darwin of The Elves" on #linode #bitlbee
17:01<alan>oh 1 sec
17:03<alan>yea still not working
17:03<alan>SELECT SUM(np_orders.paid * 0.44) As actual FROM np_affiliate JOIN np_orders ON np_orders.user_id = np_affiliate.user_id WHERE np_affiliate.ref_user_id = 2 AND np_orders.status = "paid" OR np_orders.status = "subscribed"
17:03<alan>gives 208.33
17:05<alan>https://bpaste.net/show/c1a2121f8990
17:05<alan>heres what u wante d me to pastebin
17:14<dwfreed>with that data, I get 204.81
17:14<dwfreed>by doing the whole query by hand
17:15*LouWestin is glad he’s getting help
17:15<dwfreed>now I need to go
17:21-!-fergtm [~fergtm@dyn-2001-0470-1f0f-006e-b976-59b1-d375-5300.hq-ptr.nil.com.mx] has quit [Ping timeout: 480 seconds]
17:26<alan>!point dwfreed
17:26<linbot>alan: Point given to dwfreed. (75) (Biggest fan: mcintosh, total: 17)
17:27-!-alan [~oftc-webi@107-179-235-11.cpe.teksavvy.com] has quit [Quit: Page closed]
17:28<nuevu>Not sure why he was doing math inside the sum() function.
17:29<nuevu>It's possible there was some floating point nonsense as a result.
17:38-!-fergtm [~fergtm@dyn-2001-0470-1f0f-006e-ed45-a2f9-7db1-1c35.hq-ptr.nil.com.mx] has joined #linode
17:38-!-fergtm is "Fernando" on #linode
19:35-!-eyepulp [~eyepulp@c-71-239-4-144.hsd1.il.comcast.net] has quit [Remote host closed the connection]
19:47-!-The-spiki [~spiki@0001014f.user.oftc.net] has joined #linode
19:47-!-The-spiki is "Nenad Spirkoski" on #linode
19:53-!-spiki [~spiki@0001014f.user.oftc.net] has quit [Ping timeout: 480 seconds]
20:40<chesty>what sort of field is np_orders.paid ? it's not a float is it?
20:41-!-troy [~troy@00012f14.user.oftc.net] has joined #linode
20:41-!-troy is "troy" on #oftc #linode #debian
20:48-!-thiras [~thiras@195.174.215.70] has quit [Ping timeout: 480 seconds]
20:49<dzho>whoah that is a lot of SQL in the backscroll
20:55<kharlan11>floating numbers are hard. It's more reliable to store the whole number and it's decimal in seperate fields are 4 bytes integers.
20:56<kharlan11>s/are/as
21:12<dwfreed>or just store it as an integer with a known base 10 exponent
21:12<dwfreed>this is why stripe uses pennies
21:12<dwfreed>because javascript floats suck
21:37<chesty>I don't know that floats suck if used as documented, maybe they do, but they absolutely suck if used to represent money or anything that needs accuracy.
21:39<fergtm>third time this year something breaks when doing `apt upgrade` in a Ubuntu linode
21:39<fergtm>I use stunnel to secure some connections between two linodes in different datacenters. Today there was an update for OpenSSL and it broke, lot of SSL routines:tls_construct_ctos_early_data:internal error
21:39<chesty>it's like trying to tow a caravan with a bicycle and coming to the conclusion bicycles suck. no, bicycles rock, they just suck at towing caravans
21:40<fergtm>sorry if not related to this channel, I just wanted to vent my frustration
21:41<fergtm>I expected better qc from ubuntu updates
21:42<chesty>fergtm, seems related to me. it might not be ubuntu's fault though. they might have applied a security patch from upstream.
21:45<fergtm>yeah, I'll just rollback the update and wait
21:45<chesty>fergtm, this bug looks related https://github.com/openssl/openssl/issues/7261
21:47<fergtm>it is supposed to be fixed almost a year ago. I wonder how old is the version in ubuntu repos
21:52<chesty>if that is the bug it was fixed in master and 1.1.1, what version of openssl do you have?
21:53<chesty>I have 1.1.0g
21:59-!-rnowak [~rnowak@rnowak.ombudsman.oftc.net] has quit [Quit: leaving]
22:06<fergtm>I'm typing from my phone (commuting to home) but I'll check the version later
22:11-!-squidly [~squidly@2600:3c03::f03c:91ff:fead:ed05] has quit [Remote host closed the connection]
22:43-!-Nivek [~oftc-webi@177.154.139.200] has joined #linode
22:43-!-Nivek is "OFTC WebIRC Client" on #linode
22:44<chesty>I'm on 18.04 and that bug I posted didn't seem to affect the version of openssl in it.
22:47-!-rnowak is "Robert A. Nowak" on @#oftc
22:47-!-rnowak [~rnowak@rnowak.ombudsman.oftc.net] has joined #linode
23:01<dwfreed>linbot: errno 12
23:01<linbot>dwfreed: ENOMEM (#12): Cannot allocate memory
23:01<dwfreed>oh, lol
23:03<nate>Ooooooookay, a week so far and no IPv6 issues, I feel like I'm about to let my guard down and everything will melt down again. Though I do need to figure out why lighty is having issues binding on IPv6
23:04<dwfreed>because you're binding to a specific address, and DAD hasn't completed yet by the time lighttpd tries to bind
23:07<nate>dwfreed: Nah I always have delays on the startup services and it is showing it's bound (lighty wouldn't even have started if it couldn't have), it's just rejecting. Might be something I gummer'd in the firewall recently piddling around with all the IPv6 crap
23:07<nate>This might have been broken ages ago and I just never realized or just attributed it to the "usual IPv6 buggery"
23:17-!-Ttech [~ttech@00014919.user.oftc.net] has quit [Quit: Este é o fim.]
23:26-!-Ttech [ttech@2600:3c00::17:8abf] has joined #linode
23:26-!-Ttech is "Thomas Techinus" on @#botters #moocows @#transcendence @#swift #qemu #openttd #linode #help #oftc
23:27-!-Nivek [~oftc-webi@177.154.139.200] has quit [Quit: Page closed]
---Logclosed Wed Jun 12 00:00:21 2019