[04:49] good morning [05:52] does anyone know how a script or of anyway to have iperf send an email alert when a transfer takes a certain amount of time, for example lets say sending 5GB takes < than 5 mins than email? I'm runing 16.04 and system can already send email using sendmail (emails are send when cronjobs a ran) and I'm running iperf 2.0.5 [05:53] I also use monit but can't seem to find an option for > < (greater than less than) alerts. [08:07] hey guys, I'm using ubuntu 18.04 and am trying to play with bonding. I notice that adding a bond to netplan works fine, removing a bond from netplan does not work at all (nothing happens) and setting the mtu on the physical interfaces and/or the bond does not work either [08:13] timyp: I don't think iperf has such a feature. I'd do it differently [08:14] timyp: I would place a file of X megabyte somewhere on a webserver. I'd then do: `timeout 30s wget -O /dev/null http://example.com/the_file || echo 'this took way too long' | sendmail` [08:15] good thinking [08:15] you can tinker with filesize and timeout to choose the acceptable speed. If it falls below, you'll get a mail [08:15] thank you ;) [08:23] I like the idea and configuration of netplan, I just feel it's buggy as hell [08:24] I added "match:\n macaddress: 00:11:22:33:44:55" to my ens1f1 interface, did `netplan apply` and now it's called eth0 [08:25] and interface ens1f1 is gone [08:41] possibly/probably something to do with: https://bugs.launchpad.net/ubuntu/+source/nplan/+bug/1746419 [08:41] Launchpad bug 1746419 in nplan (Ubuntu) "bond parameters are not changed by 'netplan apply'" [Undecided,Confirmed] [08:52] Hi all. I upgraded Ubuntu server from Xenial 14.04 to 18.04 yesterday and a lot of things do not work anymore. Website are broken becouse like errors like this that did not appeared before => "PHP Fatal error: Uncaught Error: Call to undefined function geoip_country_code_by_name" [08:54] Can somebody help me with this problem ? What could be the Problem ? I have installed geoip. It is in apache2 modules enabled. => "loaded Modules: geoip_module (shared)" [08:54] this should not happen ! Did Ubuntu Server downgraded geoip and deleted by this the existing working function ? [08:56] linuxperia: I'd check what php version you had on 14.04 and which one you ended up with in 18.04 [08:56] linuxperia: also check the geoip module for versioning [08:56] I think you went from 5.x to 7.y [08:56] probably there are breaking changes which you are experiencing at the moment [08:57] which is to be expected! [08:57] 18.04 != 14.04 [08:58] if you want stability and no breaking changes, stick to 14.04 [09:00] Hi TvL2386: thanks a lot for your fast answer. well there is no way back anymore! i need fix this problems asap. MySQL is broken too. Used MariaDB 10.2 before and after upgrade to 18.04 it got downgraded and now i cant install even mariaDB 10.2 baack. Its a hell at the moment [09:01] really sorry to hear that linuxperia, but this was to be expected. I'm actually really surprised that you went this path... [09:02] linuxperia: you should have tested this upgrade path before doing it [09:02] and having a rollback scenario (I like offline vm snapshots) is also very good [09:03] for mariadb, you should check out there site. There is a ubuntu ppa I think that allows you to install different versions [09:04] why ? i thinked 18.04.01 is good for upgrade. The whole Upgrade Process is for sure a joke! it deletes packages that are needed and replace it with old packages that do not work. The Upgrade process should be smart enogh to stop when it see that it want install older packages and stop with error message instead break whole working server. I really think subscribe to RedHat Debian/Ubuntu is a joke lately. Cant even install MariaDB 10.2 to use [09:04] my DataBase at the moment with 18.04 [09:04] if you can: I'd restore a backup, because you don't know how long it will take you to fix this assuming you can fix it... Probably changes in the website code are required to accomodate the new versions of php and geoip [09:05] linuxperia: packages are not replaced by old packages that do not work [09:06] yes need new programming fo the scipts. one day this hit did with MariaDB i had installed 10.2 and it deleted it with 10.1 10.2 is not existent in 18.04 [09:07] so in 14.04 you had mariadb 10.2 [09:09] yes in 14.04 i had 10.2 as i needed replication for several servers then did the ubuntu upgrade and it just installed MariaDB 10.1 and removed my Database folder /var/lib/mysql where all my data was and created a new /var/lib/mysql. At least it was smart enogh to create a copy of the existing DB before it created a new one [09:10] I have lost nearly all my nerves after this upgrade [09:10] expected be a 1 Hour work. Now its a total pain. Asking me since hours if i should switch to RedHat. i am 100% sure with redhat this would not have happened [09:11] and i am a paing Ubuntu Server Cloud Customer with monthly bills above 150 USD [09:11] aka i support cannonical [09:13] linuxperia: I'm checking at the moment, but I don't think mariadb-server 10.2 is present in 14.04 [09:14] linuxperia: I expect that testing your upgrade will take you a couple of hours [09:14] linuxperia: and I expect the actual migration to 18.04 to take days [09:14] all system packages are updated [09:15] i did installed MariaDB from the official repository website of mariadb aka used external repository. There was a message while upgrade that external repositorys will be disabled [09:15] you installed mariadb 10.2, while 14.04 comes with mariadb-server-5.5 [09:15] which means you installed from ppa [09:15] ppa for trusty [09:15] which is disabled or not working by the upgrade [09:16] and what you call "official repository website of mariadb" has nothing to do with ubuntu [09:16] es exactly but ubuntu should be smart enogh to see i have a newer DB and stop installing Old DB that is incompatible and destroy my database this for sure is a joke [09:17] well linuxperia, sorry to say it but your upgrade plan is a joke [09:17] because the plan is: there is not plan: just `os-release-upgrade` and YOLO [09:18] good luck to you sir [09:18] yes that was my plan and i dont know any other :-) [09:18] i need luck and yes lot of nerves thanks for your good wishes [09:20] hehehe... yeah you should have tested this before in a test environment... That's the other plan [09:20] the only good plan imho [09:20] and always have a foolproof rollback plan [09:20] if you have, you won't have nerves :) [09:21] btw: this advice is distro independent [09:23] yes at the end when everything is broken i got smarter now too. Should have done it the otherway around but till yet did not had such problems so expected only minor problems. well okey need solve this php geoip problem PHP Fatal error: Uncaught Error: Call to undefined function geoip_country_code_by_name [09:36] linuxperia: if you have apaid support plan, you should probably turn to coimmercial support to get help with this upgrade. [09:37] they may be more motivated to support you than us volunteers are around here [09:37] (i don't know whether commercial support is part of your current contract) [09:39] https://stackoverflow.com/questions/10821974/how-to-make-geoip-country-name-by-name-work is the first hit on a web search for "Uncaught Error: Call to undefined function geoip_country_code_by_name" [09:39] that's issue #1 of #?????? [09:40] it is common that you will need to make application level code changes when upgrading your linux distribution (and it does not matter which distribution it is) [09:40] rollback rollback [09:42] i'm unlreated to TvL2386 and, like anyone doing linux sysadmin for a couple years would probably have done, i'd told you the same had you asked before initiating the upgrade: be sure you have complete, current backups, and a rollback scenario. test and prepare for (application changes, configuration file changes, service migrations) the upgrade on (a) separate system(s) before you do it live. [09:42] this is just common practice. [09:42] ... andf do not foirget that you need to upgrade 14.04->16.04->18.04 ... else you ask for a mess (do-release-upgrade surely takes care for a lot of data and package transitions, but only between two LTS releases, you should never skip one) [09:44] when planning the next releas eupgrade (and maybe generally, too), take a look at the documentation, too, since it would also point out how to prepare for an upgrade. [09:45] i have Ubunutu Server on the Google Cloud which is a canonical product. Paying my google Cloud Bills since years using only ubuntu server to support canonical and the developers. Not sure what the commercial support by canonical is. never seen it. would be good having something like this. i did maked a search already yesterday related to this strange php error. yes its distro unrelated true. but solving it on ubuntu is really painfull. there [09:45] is no really anywhere a solution. the one is from 2015. aka nearly 3 Years old. [09:45] Which MariaDB version is in 18.04? Odd that it would downgrade [09:47] Skuggen i have instaled MariaDB 10.2 on 14.04 becouse i needed a distributed DB Replication then after upgrade of my ubuntu server to 18.04 ubuntu downgraded the DB from 10.2 to 10.1 and deleted the database as 10.2 does not exist in 18.04 [09:48] Ubuntu doesn't downgrade the package unless some other installed package depends on MariaDB in a way MariaDB's packages can't satisfy [09:50] Hm, are you sure it deleted the database? I think MariaDB packaging has some logic for copying it to a different location if it finds incompatible versions [09:50] Is there a /var/lib/mysql-10.2 or similar? [09:51] ubuntu 18.04 has mariadb 10.1 in the standard repositories [09:51] well for sure it did as i used the mariadb ppa for trusty and this ppa was disabled in bionic while the upgrade aka was non existent. think it just did mariadb-client mariadb-server upgrade and replaced the binaries then the configuration window poped up and told me that i will replace the existind db yes i have a moved db. problem is i can not even install now MariaDB 10.2 [09:51] but linuxperia has used the mariadb ubuntu ppa to get 10.2 on his system [09:57] maybe just use their ppa again then? [09:58] linuxperia: how can you even think about uppgrading a OS without full backup and rollback plan? [10:03] tomrey: trying since hours to use their ppa for bionic but for some strange reason i cant install MariaDB 10.2 anymore on 18.04 It brings this error message here always! "sudo apt-get install mariadb-client-10.2 mariadb-client-core-10.2 mariadb-server-10.2 mariadb-server-core-10.2 Für Paket »mariadb-server-10.2« existiert kein Installationskandidat" [10:05] its total strange. apt see the package but it cant install it. i dont know what to do anymore. this problems should not happen. i follow all the steps add the ppa do apt updata and then a simple apt install but it just wont install anymore my existing MariaDB 10.2 on 18.04 [10:08] here is the Paste Bin of the Output => https://paste.ubuntu.com/p/qWJ94p9kmC/ [10:20] linuxperia: this is the commercial support offering i had was discussing https://www.ubuntu.com/support [10:23] mariadb (since you're using their software repository) also offers commercial support. [10:27] yes i even donated money to them. will need check out their offer. think they can not help me however when ubuntu refuses to install the MariaDB 10.02 on 18.04 as this is ubuntu Server related and not really MariaSB related. its strange the i get this error whenever i try install MariaDB on 18.04 [10:27] The following packages have unmet dependencies: [10:27] mariadb-server : Depends: mariadb-server-10.2 (>= 1:10.2.17+maria~bionic) but it is not installable [10:27] E: Unable to correct problems, you have held broken packages. [10:28] i dont have any broken packages however [10:28] and i follwed thos official install guide here https://downloads.mariadb.org/mariadb/repositories/#mirror=host-europe&version=10.2 [10:28] thanks for the support link will go check it out [10:34] tomreeyn: looks like that support is only availble for main packages. not sure what is possible at all with such a limites offer. Support for MariaDB 10.2 on ubuntu 18.04 installation will be clearly not covered by this offer [10:37] linuxperia: which is why i suggested you also look into mariadb support, if you prefer using their packages. [10:39] note that thereis also #mariadb aroun dhere, i think [10:40] yes thank you very much tomreyn. i appreciate your help effort. [10:41] linuxperia: you're welcome, good luck. [12:01] good morning [12:19] Was thinking to myself on if it makes more sense to have a folder location binded instead of syslinked [12:22] Kinda makes sense in my head. You'd have all your links defined in the /etc/fstab [12:23] Instead of checking each directory and inspecting if a symlink was used. [12:26] and maybe it's just yak shaving. it probably depends on your needs / use case. [12:27] Yak shaving? [12:27] https://en.wiktionary.org/wiki/yak_shaving [12:28] And I suspect it'd make the most sense for important folders like /var if it's really somewhere else on the system. Fails a bit more gracefully with the system entering single user mode. [12:29] Ah, yeah I suppose it's me thinking about useless stuff. [12:34] Well, something less useless: What happens if a formated block device expands? Think /dev/my-block-device formated as ext4 has expanded 1gb. Does it just work and the space is instantly usable or does stuff break? [12:39] CarlenWhite: resizing block devices is not usually the same as resizing file systems, althought there are some utilities which will / can do this in just one go. [12:42] That's what I figured. I'm messing with zfs's zvols which gives you block devices and they can be resized, but I've been bubblewrapping the data by partitioning. [12:46] hm, snapper upstream: "Latest commit d610a3a on Jul 25" [12:47] not promising [12:58] does it do ext4 by now? [12:59] tomreyn: they say this in the manpage: "Supported filesystems are btrfs and ext4 as well as snapshots of LVM logical [12:59] volumes with thin-provisioning. Some filesystems might not be supported depending on your installation. [12:59] " [13:00] thanks ahasenack, so i guess their faq is outdated: http://snapper.io/faq.html -> "Does snapper support ext4? - Yes, but only experimentally and you need a special kernel and e2fsprogs. For more information see the next4 project." [13:01] well, don't know [13:01] ext4 is not something I would think of when dealing with snapshots [13:01] you use xfs? [13:02] last I checked, xfs can't do snaps either [13:02] or lvm with thin provisioning? i think i spotted you discuss LVM2 snapshots on 'normal' LVs vs thin provisioned LVs a while ago. [13:02] nor anything else, except btrfs and zfs [13:03] and lvm of course, but that's dead slow [13:03] btrfs is ok, for a while, but only just [13:03] zfs just works [13:04] RoyK: i think snapper isn't meant to control snapshotting for file systems which already support it but for doing it on file systems which dont do it natively. [13:04] tomreyn: zfs [13:05] i see [13:05] but I don't use snapper [13:05] I'm just fixing a bug there [13:05] it's failing to build in ubuntu [13:07] oh, and probably on a current ubuntu release, which the travis ci builds wouldn't know about. [13:08] hmm their travis ci builds run on 17.10 [13:09] I submitted a patch [13:09] https://github.com/openSUSE/snapper/issues/424 [13:09] no idea if someone is watching that, though [13:09] given the last commit was months ago [13:21] here it's Sep 3, which is less than 6 weeks since Jul 25 [13:22] tomreyn: seems it controls snapshotting for lvm and btrfs [13:29] cpaelzer: is the importer still catching up? [13:43] ahasenack: yep [13:43] still running [13:43] let me scroll up to get a feeling on the overall load [13:44] I see this known issue passing by "User is suspended: ~python-modules-team'"" [13:45] ahasenack: 59 done, there is no preview how much more will follow [13:46] ok [13:50] never heard of snapper, but reading this: http://snapper.io/overview.html makes me very much interested [13:50] snapshots of ext4 filesystems???????? [13:51] what are the caveats [13:51] snapshot ; apt-get update && apt-get dist-upgrade -y ; and rollback if it failed???????????? [13:52] TvL2386: sounds strange - ext4 doesn't have any support for snaps [13:52] RoyK: yeah I totally agree. I'm really curious how they have implemented it [13:52] gonna check the source [13:57] TvL2386: I guess lvm snaps [13:57] and some hooks in ext4 to sync before the snapshot is taken [14:21] openvpn is such a pain to setup :/ [14:23] ahasenack: yep [14:23] I'll optimize the steps a bit [14:23] cpaelzer: have you seen this? https://github.com/FiloSottile/mkcert [14:23] looks easier than easy-rsa [14:25] I tought I have read about it, but find nothing in my history [14:25] maybe the name is similar to something else I touched [14:35] wireguard to the rescue! [14:41] ahasenack: given what flies by in the log there are plenty of packages that are prevented to import by that 410 error of the deleted team [14:41] :/ [14:41] and the fix is to just ignore that error [14:43] cpaelzer: I'm fetching pkg every 5min or so, still no update for snapper [14:43] the backlog must be big [14:43] yep [14:43] it has all of the "rush into cosmic" phase [14:44] you might remember the discussion when I asked if rbasak usually keeps the log open to check [14:44] and the answer was no so I did the same [14:44] it seems two days after it failed hard enough to stop [14:44] there are import fails (fine) but the bad ones are those which kill the watch/loop to import [14:45] this definitely has to change [14:45] not you looking at it daily, no :) [14:45] yeah, including a trivial status page [14:45] it should become a real service [14:45] as we depend on it [14:45] like published on LP on colum #1, imports on column #2 with status [14:57] cpaelzer: hm, snapper is not in the whitelist, it won't get (re)imported automatically [14:57] cpaelzer: can you run it on the side? [14:58] heh [14:58] yeah [15:07] cpaelzer: got it updated, thanks [15:12] yw [15:12] ahasenack: I was still wondering why it touched so much [15:13] I saw like 20 versions import [15:13] but I'd have expected only 1 new [15:13] ahasenack: when you fetched that did it mark all branches as forced-update? [15:13] no [15:13] due to any import accident? [15:13] hmm [15:14] maybe it has to pass all old upload always [15:14] cpaelzer: this was the output: https://pastebin.ubuntu.com/p/35SqT62v3S/ [15:14] I remember nacc takling about noting somewhere where to pick up next time to speed things up [15:14] it was manually imported for the first time last week [15:14] could be related [15:14] it's in universe, so we didn't have it yet [15:15] but last week to this week should not be that much IMHO ahasenack: http://paste.ubuntu.com/p/scfMKBbw7k/ [15:15] no clue [15:15] I'd expect like 1x fetch what it already has - then jumping to the only new one from the publish history [15:15] I did not run --no-fetch on this [15:16] ahasenack: if you added a topic to the sprint you might add this pastebin as well [15:16] I can, sure [15:21] fresh install of 18.04, apparently I'm using netplan. I have to log in and sudo netplan apply for my network to work after a reboot, what am I doing wrong [15:36] mybalzitch: does syslog tell? [15:36] https://askubuntu.com/questions/1019146/netplan-does-not-apply-at-startup I think its that [15:36] and the systemd-networkd logs stop after it renames the interface [15:36] did you install 18.04.0 or 18.04.1? [15:37] 18.04.01 [15:39] do you have multiple NICs then? [15:39] https://0bin.net/paste/h37PDNKKR3DoTGGM#mdk6bzkx9WjCGGmAVj7U4d1WJnxC9iRyuDa4NuTez7i [15:40] no, just the one virtio device [15:40] but I use vlans [15:42] mybalzitch: hmm, i'd say file a bug, using "ubuntu-bug systemd" [15:43] mybalzitch: just to make sure: you had run 'netplan apply' after changing the netplan configuration and before rebooting, right? [15:56] you mean netplan generate ? [16:25] either way, yes I've run netplan apply, that's the only way to get the network working [19:10] mybalzitch: i was asking whether you ran "netplan apply" (or "netplan generate") before you rebooted, to ensure the systemd-networkd configuration is updated for next reboot. since if it wasn't, i would think the previous configuration would have applied by the time you rebooted. [19:18] Greetings all, if any of you have firewall/router experience and would like t help me out. I've placed my issue here: https://pastebin.com/Z5mRQSjQ [19:54] Hey there. Given that I have a service running as someuser, what is the proper way to have that service take on the same environment parameters as if you were logged in interactively? I have a work around by using Environment="PATH=foo:bar:baz". But that means I would have to change every custom unit file whenever something is added to the users env (or even when just one env value changes). [20:45] Hello guys! [20:46] I am looking for a error log from mysql on ubuntu server which should be present at /var/log/mysql/error.log but the file is clear. So I wonder where is the general log file from ubuntu server please? [20:50] supercool: it might change by ubuntu release and based on the mysql server variant you're using, but i would indeed expect to see it in this directory (the file might also be called 'mysql-error.log', though) [20:51] supercool: if mysql fails to startup properly then logging to /var/log wont be possible and it will instead log to its data directory, so (by default) to /var/lib/mysql/(data/) [23:00] Thank you tomreyn ! [23:56] Could someone help me figure why mysql-server is not starting here please? [23:58] paste your error [23:58] ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) [23:59] supercool: did you find the mysql error log? [23:59] that line above just means mysql is not running [23:59] ahasenack: not really. I fould some logs but nothing helpped.