[05:40] <cpaelzer> good morning
[05:58] <lordievader> Good morning
[07:34] <sonu_nk> i have a linode server . installed ubuntu 14 .. can i upgrade it to ubutnu 16 without remove any ubuntu 14 setting website and other data ?
[08:49] <rh10> guys, how can i send beep sound from script? for localhost using
[08:49] <rh10> i need notify sound when script is done
[09:08] <rh10> got it. now im using paplay from pulseaudio
[09:08] <rh10> paplay /path/to/file
[09:24] <rh10> but! it doesnt play sound when  it runs from cron
[09:24] <rh10> all path absolute
[09:25] <rh10> i tried run from user cron, from /etc/crontab - with username added
[09:31] <FMan1988> maybe it does not have permission to access the file
[09:32] <rh10> FMan1988, i checked it. all correctly
[09:33] <rh10> plays normal when i run manually, and dont play when it runs from cron
[09:34] <rh10> * when i run script manually within paplay run
[12:14] <jamespage> coreycb: gosh - tl;dr - don't change /usr/bin/python - it breaks stuff
[12:15] <jamespage> I must have triaged 6 bugs related to users switching that link to python3 today
[12:15] <coreycb> jamespage: oh no
[12:16] <jamespage> coreycb: yeah the debpython module that gets called from maintainer scripts for python modules is py2 only @16.04
[12:16] <jamespage> if you see a bug with an unable to import from ConfigParser,its this type of problem
[12:16] <jamespage> typically on a package install
[12:16] <coreycb> jamespage: ok good to know, thanks
[13:16] <albech> is it possible to roll back one upgrade? The latest postfix upgrade totally hosed my system.
[13:32] <Bert_2> Hi guys, can I allow-hotplug vlan interfaces such as "eth0.20"?
[13:45] <smoser> albech: you can probably:
[13:45] <smoser>  apt-get install postfix=3.1.0-3
[13:45] <smoser> albech: could you please file a bug describing your regression though?
[14:02] <albech> smoser: was 3.1.0-3 the previous version?
[14:07] <albech> smoser: looks like it will lead to some dependency issues with postfix-mysql and possibly others..
[14:07] <albech> smoser: at least i can receive mails at the moment, they just arent passing through rspamd :(
[14:10] <sdeziel> albech: yes 3.1.0-3 is the previous version
[14:11] <sdeziel> albech: during the upgrade, did you get a prompt about a changed master.cf file?
[14:19] <smoser> albech: yes, it was the previous version.
[14:21] <albech> sdeziel: nope it said it would keep the old version
[14:22] <smoser> nacc: when your'e in... i had a question on https://bugs.launchpad.net/maas/+bug/1444992
[14:22] <ubottu> Launchpad bug 1444992 in MAAS "fastpath install duplicates iSCSI initiator names, blocking iSCSI HW" [High,Triaged]
[14:22] <albech> getting lots of these (for every mail connection): https://pastebin.com/Aav8gwyu
[14:24] <sdeziel> albech: can you compare the current master.cf and main.cf with those from before the upgrade?
[14:26] <albech> sdeziel: they werent changed
[14:26] <sdeziel> I don't know how it could have break your rspamd integration then, sorry
[14:27] <albech> sdeziel: not even sure what that SMFIC_OPTNEG means
[14:28] <albech> but its a real issue. letting roughly 1000 spam/minute through atm :/
[14:29] <sdeziel> albech: it seems to be how postfix negotiates feature support with the rspamd milter
[14:29] <albech> maybe a little less 900ish
[14:29] <sdeziel> albech: could you try a manual connection yourself: nc -zv localhost 11332
[14:30] <sdeziel> cause the connection timeout is weird
[14:30] <albech> Connection to localhost 11332 port [tcp/*] succeeded!
[14:33] <sdeziel> albech: try to check if the chroot version has the same localhost: grep localhost /var/spool/postfix/etc/hosts
[14:34] <sdeziel> albech: then try connecting with IPv4 then IPv6: nc -zv4 or nc -zv6
[14:34] <sdeziel> albech: anything useful logged in /var/log/mail.err ?
[14:35] <albech> mail.err has another error related to the upgrade i guess: Oct 24 16:21:09 maia dovecot: imap-login: Error: SSL: Stacked error: error:14094416:SSL routines:ssl3_read_bytes:sslv3 alert certificate unknown: SSL alert number 46
[14:36] <sdeziel> OK, that's unrelated
[14:36] <albech> looks fine https://pastebin.com/HkvnUqbh
[14:36] <sdeziel> albech: I never used rspamd but I'd check it's logs and maybe restart it?
[14:37] <albech> tried all.. logs dont show anything. i have even turned on verbose logging for the milter plugin
[14:38] <albech> nc connects fine with both v4 and v6
[14:40] <sdeziel> albech: I'd try to tcpdump lo and see the milter dialogue between postfix and TCP/11332
[14:40] <albech> did that also, but its above my skill level to read the output
[14:41] <albech> sdeziel: maybe too verbose.. let me know: https://pastebin.com/KecaXqEK
[14:42] <albech> i could be wrong, but these are all ipv6 requests?
[14:42] <sdeziel> looks like it
[14:43] <albech> wondering why.. never told it to use ipv6
[14:43] <sdeziel> IPv6 is used by default when available
[14:47] <sdeziel> albech: did you try rolling back to 3.1.0-3 as suggested? You can do so for all the postfix-* packages
[14:48] <albech> sdeziel: no, cause they started some dependency issues with postfix-mysql and possibly other packages.
[14:48] <albech> sdeziel: think its worth a try?
[14:49] <sdeziel> I'd try it if I was hammered by 900 spams/minutes :)
[14:50] <albech> i prefer the spam over a non functional system. tried getting into dependency hell before. i was just stupid not doing a snapshot before the upgrade.. note to self :/
[14:52] <albech> how could i grab the right version of postfix-mysql along with the 3.1.0-3?
[14:53] <albech> think i figured that out ;)
[14:54] <sdeziel> apt-get install postfix{,-mysql}=3.1.0-3
[14:55] <albech> sdeziel: got that
[14:56] <nacc> smoser: pong
[14:59] <albech> sdeziel: after the downgrade the problem is still there. the pastebin is the upgrade done that caused the issue in the first place. Wondering which other packages could cause it. https://pastebin.com/p45YnyRa
[15:00] <sdeziel> albech: rspamd:amd64 (1.6.4-1~xenial, 1.6.5-1~xenial) expected?
[15:00] <albech> sdeziel: hmm yes i think so ;)
[15:01] <albech> sdeziel: not sure why it shouldnt be
[15:02] <sdeziel> albech: your milter communication is broken so it's either the initiator (postfix) or the responder (rspamd)
[15:02] <albech> sdeziel: aye
[15:02] <albech> could try to downgrade the rspamd then
[15:03] <smoser> nacc: hm... so what happened there... we just re-generate that id every boot ?
[15:03] <nacc> smoser: it should regenerate on first boot
[15:03] <albech> sdeziel: how can i tell what the previous version was?
[15:03] <smoser> what stops it from happening every boot ?
[15:04] <sdeziel> albech: apt-get install rpamd:amd64=1.6.4-1~xenial (see the /var/log/apt/history.log you pasted)
[15:05] <nacc> smoser: /etc/init.d/iscsid itself (which calls /lib/open-iscsi/startup-checks.sh)
[15:05] <nacc> smoser: and overwrites GenerateName=yes with a real id
[15:09] <sdeziel> albech: is it working better now?
[15:09] <smoser> nacc: ok. i trust. wasnt clear in my memory.
[15:09] <albech> sdeziel: E: Version '1.6.4-1~xenial' for 'rspamd' was not found
[15:10] <albech> trying to get the old version
[15:10] <sdeziel> albech: maybe the old package is still available in the PPA, please paste "apt-cache policy rspamd"
[15:11] <nacc> smoser: so, in theory, if you d/l the image and examine it without running it, it should have GenerateName=yes in it. But if you launch an image, when iscsid starts it will generated a new IntiatorName value and store it there. That image will keep that for its life. If you snapshot it and launch it again from there, it should keep it too. But if you spawn a new one from the cloud image, it should get
[15:11] <nacc> a new Name
[15:11] <nacc> smoser: that was our test in devel, and what looks to be the result in xenial
[15:11] <albech> sdeziel: doesnt look like it: https://pastebin.com/WZ0KfMiU
[15:13] <sdeziel> albech: check in /var/cache/apt/archives
[15:14] <albech> sdeziel: no only rspamd_1.6.5-1~xenial_amd64.deb
[15:15] <sdeziel> albech: they've made many milter related changes according to https://github.com/vstakhov/rspamd/releases/tag/1.6.5 so it really looks like a regression
[15:15] <albech> sdeziel: :(
[15:16] <sdeziel> albech: with 1.6.4 but looks similar https://github.com/vstakhov/rspamd/issues/1869
[15:17] <sdeziel> albech: at this point I think you should hop in #rspamd
[15:18] <albech> sdeziel: i am trying.. thanks for your help.. appreciate it
[15:18] <sdeziel> np
[15:57] <rbasak> nacc: can you happen to think of a good example package for considering the commit graph change?
[16:31] <teward> i owe the server team an apology today.
[16:31] <teward> there was an emergency that took my attention - a close friend of mine got in a car wreck and i needed to make sure they were okay.  (I was at the hospital all morning...)
[16:31] <teward> my apologies for not being here or letting anyone know, i only found out about the accident this morning so...
[16:32] <dpb1> ugh
[16:32] <dpb1> I'm sorry teward, best wishes to your friend
[16:32] <teward> dpb1: tell me about it
[16:32] <teward> dpb1: they're a bit screwed up right now, and not entirely there in the head right now due to the morphine, but they're alive.
[16:33] <dpb1> :/
[16:33] <teward> broken bones, lots of cuts and bruises, and a major concussion
[16:33] <dpb1> yuck, lucky to be alive indeed!  sounds like a major accident!
[16:33] <teward> i'd have let someone know but i kinda left everything at my apartment when I bolted out the door this morning...
[16:33] <teward> ... ended up leaving all of my devices at home.
[16:33] <teward> dpb1: sedan vs. truck.  yeah pretty major.
[16:33] <teward> they were in the sedan
[16:34] <dpb1> teward: no worries at all.  an irc meeting is very insignificant
[16:39] <powersj> teward: no apology needed
[16:39] <powersj> teward: I did put you down again for next week, so let me know if I should change that
[16:40] <teward> powersj: you can leave me listed there for now.  i'll try and let you know if I can't make it.
[17:00] <rbasak> teward: :-(
[17:00] <rbasak> Best wishes to your friend, and no apology needed :)
[17:01] <sudormrf> question about MDADM. I have a 4 bay case with 4 drives. I am going to set up RAID10, but I have data on one of the drives. What I want to do is this: create an array out of 2 disks, copy the data over to that array, then add 2 more disks to it resulting in RAID10. Is this possible to do without a 5th drive to first copy the data to?
[17:01] <sudormrf> I believe it should be
[17:02] <teward> sudormrf: to do the RAID array will nuke the disk data.  you'll need to back up the data you want to keep first, if I'm not mistaken.
[17:02] <sudormrf> teward: to get it to a RAID10 state, you mean. correct?
[17:03] <teward> sudormrf: is it currently RAIDed?
[17:03] <sudormrf> negative
[17:03] <sudormrf> JBOD right now
[17:04] <teward> sudormrf: in *theory*, maybe.  But I'm not as well versed in RAID10; in theory you can create the array with the other disks, and leave it in a degraded state by marking the first disk as missing, then copy the data onto the array, then extend the array onto the original first disk.  But, I can't guarantee you wont lose data this way either
[17:05] <teward> so you should probably do a backup first.
[17:05] <sudormrf> hmmm
[17:22] <nacc> rbasak: off the top of my head -- no, but anything that successfully imports should work fine. ipsec-tools might be reasonably quick
[17:23] <dpb1> sudormrf: I think the general advice is to rebuild from scratch as raid10 and copy your data.  But, It does appear to be possible to go from raid0 to raid10 now.  So, I suspect you could do as you are expecting.  I would build a VM and try it out first.
[17:23] <sdeziel> sudormrf: you should be able to create your array in degraded mode (with one drive missing)
[17:24] <sudormrf> sdeziel: hmm. ok.
[17:24] <sudormrf> dpb1: I will give it a go.
[17:24] <sdeziel> sudormrf: so if you put all your data on one drive, create the RAID10 with the other 3 drives you should then be able to copy the data to the RAID then recycle the single drive by joining it into the array
[17:24] <sudormrf> I've only built em from scratch before, so this is a bit of uncharted territory for me
[17:24] <sudormrf> rad
[17:35] <teward> quick check with the server team:
[17:35] <teward> with the upcoming cycle approaching LTS, should I be focusing on nginx mainline in B-series, like we did for Xenial with a post-release targeted update to the stable version
[17:35] <teward> if there's no objections :p
[18:04] <sdeziel> sudosmurf: I just tested this with loop devices and it works well :)
[18:07] <sudosmurf> sdeziel: build array in degraded status?
[18:07] <sudosmurf> so you just give it 3 device names instead of 4?
[18:08] <sdeziel> sudosmurf: yes, built it with: mdadm --create /dev/md3 --run --level=10 --chunk=4 --raid-devices=4 /dev/loop{0,1,2} missing
[18:08] <sdeziel> sudosmurf: then later on, I added loop3 to the mix
[18:10] <sudosmurf> rad. what are loop devices, BTW?
[18:11] <Walex> sdeziel: building RAID5 and especially RAID6 sets in degraded status is recommended as it can be much faster
[18:11] <Walex> sdeziel: in the case of RAID5 it is actually sort of automatic
[18:12] <sdeziel> Walex: hmm, didn't know that, thx
[18:12] <sudosmurf> Walex: good info :). In this case the array is going to be RAID10.
[18:12] <sdeziel> sudosmurf: loop devices are devices backed by files to simulate a disk. In other words, I didn't want to test with 4 physical disks so I used plain files instead
[18:14] <sudosmurf> oh neat
[18:15] <sudosmurf> TIL :D
[18:23] <JanC> "recommended" depends on how you consider the risks of running without redundancy for as long as it takes to sync the redundant disk...
[18:25] <JanC> sdeziel: you can also use in-memory block devices for testing instead
[18:40] <sdeziel> JanC: interesting TIL about the brd module
[19:19] <albech> been having some problems after an upgrade and notice that the following packages has been kept back: python-cryptography python-openssl python-requests python-urllib3 python3-requests python3-urllib3. Could it possibly be related to this? Can i force the install of these packages?
[19:19] <sdeziel> albech: it would be pretty surprising if those were involved with the milter problem that you faced earlier
[19:20] <sarnold> something like apt-get install python-cryptography python-openssl python-requests     ought to explain -why- they were held
[19:20] <albech> sdeziel: dont know how much of rspamd is written in python.
[19:21] <sdeziel> albech: any success in rolling back rspamd itself?
[19:21] <albech> sdeziel: none
[19:22] <albech> sdeziel: #rspamd is pretty quiet about the issue.. i cannot be the only one facing it
[19:22] <albech> sdeziel: its a pretty standard build i have.
[19:22] <sdeziel> albech: can you paste "apt-cache show rpamd"
[19:22] <albech> sdeziel: but at least you gave me some good tips on the trouble shooting which i have forwarded to the channel.
[19:23] <albech> https://pastebin.com/dr7iyw3n
[19:24] <sdeziel> albech: this rspamd package not being provided by Ubuntu/Canonical means that you depend on rspamd upstream to help you with it
[19:24] <albech> sdeziel: aye
[19:24] <sdeziel> albech: none of the dependencies are related to python
[19:25] <albech> sdeziel: dont think its in the ubuntu repo at all
[19:25] <sdeziel> albech: it's not indeed
[19:25] <sdeziel> albech: libicu55 was updated yesterday (security), so could be a regression in there, dunno
[19:26] <albech> sdeziel: ok. i will pass that on to #rspamd for troubleshooting.. thanks
[19:26] <sdeziel> albech: that said, my money would be on a regression with rspamd's new version
[19:26] <albech> sdeziel: 8:15:03 PM - cebka: albech: I doubt it is related to 1.6.5
[19:26] <albech> sdeziel: guess he should say that ;)
[21:12] <albech> sdeziel: not much help from #rspamd except: 9:36:57 PM - cebka: IDK 9:37:05 PM - cebka: your system is definitely fucked up.
[21:12] <sdeziel> albech: that's a helpful remark ;)
[21:12] <albech> sdeziel: very constructive ;)
[21:14] <albech> sdeziel: when trying to access rspamd's webinterface i am getting the following: https://paste.ubuntu.com/25812383/
[21:14] <albech> sdeziel: so its not only the milter
[21:15] <albech> sdeziel: i cannot read much from that tcpdump
[21:16] <sdeziel> albech: looks like your GET goes unreplied
[21:18] <albech> sdeziel: but at least it tells me that one more of the rspamd components arent working as expected.
[21:20] <sdeziel> albech: you could strace the rspamd process and see what it's doing
[21:21] <albech> sdeziel: let me try
[21:21] <sdeziel> albech: but before doing that, have you been able to get some help to access rspamd's logs?
[21:22] <albech> sdeziel: still only very little in the logs and #rspamd said it was expected stuff.
[21:23] <sdeziel> albech: this is IMHO off topic since rspamd isn't even shipped in Ubuntu so to not annoy folks in here, I propose we move to a private conversation
[21:23] <albech> sdeziel: fair enough, sorry to the #