[07:02] hello all. [08:02] Hello [08:07] Hellos, someone help me with mdadm problem, i want to add a disk to my new mdadm array but it run so slow, and the system freeze... https://hastebin.com/etivovonar.sql === JanC is now known as Guest97495 === JanC_ is now known as JanC === Guest76180 is now known as FreddG [12:01] Heya guys, Is there any good guide on how to create a self-signed certificate for TLS use? [12:02] NwS: Self-signed as in directly self-signed, or as in self-signed using your own CA? [12:03] andol, I have no idea tbh.. -.-" I just want to make the mails use TLS instead of nothing atm.. [12:04] NwS: Might as well grab a "real" certificate from Let's Encrypt then? [12:06] andol, As far as I remember that only lasts for a few months or something? [12:09] NwS: Yepp, hence the automated renewal [12:09] Ok will check it out. Thanks! === Piper-Off is now known as Monthrect [12:47] NwS: most mail servers will reject self signed certs [12:47] keep that in mind [12:47] Thank you for the info ikonia! === devil is now known as Guest95140 === Guest95140 is now known as devil_ === devil_ is now known as devil__ === devil__ is now known as devil === devil is now known as devil_ === Metacity|uh-oh is now known as Metacity === devil is now known as Guest69784 [15:12] MASM_: the reason why your RAID array (in a mirroring configuration) goes slower when you add a new member is that data is synched initially. See line 18 for the initiated synjch process. Once this is complete, operation should be as fast as before or faster. [15:14] oh okay, i missed the very low synch speed of 541K/sec [15:14] tomreyn: and what i need to do?, it is so slow, 200 k/s [15:15] tomreyn: right now it was 94k/s [15:16] yes this suggests something is not right - unless you have ongoing massive reads / writes? [15:18] tomreyn: the server is in runtime, but when i stop mysql, apache, it was the same... [15:18] bad partition alignment could cause reduced performance, but usually not that much. [15:18] tomreyn: You recomend me that run server in recovery mode and do it?, [15:20] MASM_: if you removed the production workload off the disks and the performance remained in the KB/s area then this won't help. [15:21] MASM_: but you can try the laignment check: "sudo parted /dev/sdX align-check optimal N" where X is the block device (such as "sda") and N is the partition number. [15:23] tomreyn: 3 partitions said aligned [15:26] tomreyn: from actual disk and new disk, i start a check in new disck with smartclt, maybe the new disk is failing [15:28] MASM_: yes, you can run a -t long on those [15:28] MASM_: are sdb and sdc the same disk model? [15:28] which disk models are those anyways? [15:29] tomreyn: who i find it? [15:32] sudo hdparm -i /dev/sd{a,b,c} [15:34] tomreyn: new disk is ST1000DM003-1SB102 , second ST1000DM003-1CH162 [15:34] "sudo smartctl --info /dev/sdX" provides more readable output [15:36] tomreyn: https://hastebin.com/widimelihi.vbs [15:36] tomreyn: https://hastebin.com/vigumokufe.vbs [15:37] tomreyn: https://hastebin.com/biquxedike.cs [15:39] so those are seagate barracudas 1TB, different generations apparently [15:39] i'm trying to understand what the difference between 1SB102 and 1CH162 is [15:45] tomreyn: i think its the generation of disk...¿?¿? [15:46] MASM_: yes, probably, it's the same marketing model, just different firmware versions. have you looked at the firmware updates information? [15:50] tomreyn: no, I intuited [15:50] tomreyn: i was searching that, i only found a pdf with first disk, the new [15:51] MASM_: so read those links which are found in the smartctl -i output you posted [15:51] (or rather the web pages these links point to) [15:54] is there an easy way in Postfix to tell it to deliver mail for a given email address to a specific user on the system? [15:54] maybe in aliases somewhere? [15:55] teward: http://serverfault.com/questions/144325/how-to-redirect-all-postfix-emails-to-one-external-email-address [15:55] teward: virtual mail domains or just a single (system) domain? [15:57] tomreyn: ultimately irrelevant, because all domains on this box would point to the same user(s). [15:57] tomreyn: though, if we're being specific, for now a single system domain [15:58] tomreyn: as in, mail to foo@bar.baz and foo@bar.baz.bash would both to to user foobar instead fo 'foo' [15:58] teward: well then just add an alias mapping, rad aliases (5) [15:58] basically, just trying to map a specific address to a specific local user mailbox [15:58] thanks [15:58] i.e. you probably will just end up editing /etc/aliases or /etc/postfix/aliases and add somethign along the lines of: [15:59] I sometimes forget my aliases knowledge :) [15:59] originalrecipient: newrecipient@example.org [15:59] or, in your case: [16:00] originallocaluser: newlocaluser [16:01] my postfix is rusty, though, i hope that's correct OTOH. [16:02] MASM_: in case you're not happy to do the firmware updates you may be able to have the disks replaced instead. also, did the long self tests complete, yet? [16:08] tomreyn: i see it will last 108 minutes, it finish in 45 minutes [16:09] tomreyn: i have server dedicate, i saw one disk is failing, mdadm raid send me mails about it, i send to hosting, but i think they didn't change it... [16:11] MASM_: you should note down serial numbers when this happens [16:11] also "sudo smartctl -a /dev/sdX" may tell [16:12] and surely dmesg -T [16:12] if a disk is broken this could very well explain the bad performance you see [16:18] tomreyn: i was searching about the ids i take noete, yes they change it [16:20] tomreyn: i need wait to finish long test in /dev/sda for do a "sudo smartctl -a /dev/sdX" [16:26] MASM_: you should already be able to run it now, getting the latest information [16:30] tomreyn: this is the result https://hastebin.com/uxusopabux.erl [16:33] MASM_: thats the ST1000DM003-1SB102 - your newest HDD, i guess [16:34] it's only been powered on for 65 hours [16:34] what about the other two (? I saw you had sda, sdb, sdc) disks? [16:36] sorry i didn't say you, when i start i saw /dev/sdb and /dev/sdc [16:36] tomreyn: and then when i reboot the server the /dev/sdc change to /dev/sda [16:38] tomreyn: and yes /dev/sda was /dev/sdc, and it is the new disk, i confirm the device serial number [16:38] that's why you should always use /dev/disk/by-id in scripts and config files [16:39] the other names are fine for humans who know they change [16:41] i see. so what's the output of "sudo smartctl -a /dev/disk/by-id/ata-ST1000DM003-1CH162_S1D9YSS6" [16:41] qman: thank for the information, I did not know it, i'm a beginner in this [16:42] that's probably sdb now [16:42] and are there any ata errors in dmesg [16:43] tomreyn: https://hastebin.com/iluvizafob.erl [16:46] okay, so both disks look alright for now, except the possible need for a firmware upgrade, except the unfinished long offline test. [16:49] there were, however, ATA ABRT errors on ST1000DM003-1SB102_Z9A4SRJX when it had 58 hours of uptime (currently 7 hours uptime ago). [16:50] this could point to bad wires, connectors, disk issues, controller issues. [16:51] also power supply and thermal issues, but less likely. [16:53] tomreyn: maybe because i reboot server, with panel from hosting, because it was freezing, like when they are resync :S <-- Could this have been? [18:13] MASM_: no, rebooting the server should not normally result in aborted ATA commands, [18:15] tomreyn: okey, i'm waiting for test it have 50% of it. , changin conversation, i remember that before some one help me and i think that person was you, 3 or 4 months ago, xD [18:16] i wouldnt remember [18:17] tomreyn: don't remember this :http://askubuntu.com/questions/809823/error-raid1-not-unsynchronized-a-degradedarray-event-had-been-detected-on-md-d [18:20] i'm afraid i don't [18:21] but my memory is not very good. [18:22] tomreyn: don't worry, your memory is selective [18:31] ok ;) === Guest69784 is now known as devil_ === voldemort is now known as nchambers === JanC_ is now known as JanC [21:28] tomreyn: https://hastebin.com/orocenocab.sql this is the new information about smartctl check with long test [21:31] MASM: one of them hadn't finished, yet: "# 1 Extended offline Self-test routine in progress 10% 70 -" [21:31] 10% lef to be checked [21:32] the other jdd looks fine, though [21:32] *hdd [21:32] hello. [21:33] I found this irc channel from the conjure-up.io page [21:33] Is this a room that provides guidance with openstack deployment? [21:35] I have been trying to get a POC off the ground. I have a beefy 56core/256ram physical box with vmware running. [21:35] I have been deploying 16.04 over and over trying to use the conjure-up method but I keep running into issue. The first was the accounts.yam file missing. [21:36] I was able to get past that and now I had around 5 of my vm instances running, then it bombed out due to a "handoff error" [21:48] So is that a no on the assistance? [22:32] tomreyn: https://hastebin.com/ipazeteguv.sql it finish, it isn't with errors [22:35] tomreyn it will be the firmware ? :S i never change firmware of hard disk [22:36] MASM: might be, might not be. could also be the controller, or bios, or something else entirely. you shouldn't do it yourself if this is a rented server [22:36] at least not before checking with support [22:37] tomreyn: yes it is a rented server [22:43] tomreyn: the slow resync will be because i didn't use " fdisk /dev/sda " to format disk? [23:11] MASM: the sfdisk export and reimport method should be fine for duplicating the partition table [23:12] MASM: check the output of "blkid" for duplicate id's, just in case [23:13] this shouldn't have created any, though, since partition tables and partitioned block devices don't have uuids assigned, just the partitions themselves. [23:20] what is the purpose of the "ubuntu" user? I've started using my first VPS and was surprised to find a user under this name. [23:21] adrian_1908: this will be specific to the images your host provides [23:21] it is not a default user account [23:21] tomreyn: ah ok, thanks. I'm using OVH. [23:21] you could try asking in the unofficial #ovh channel then [23:22] I didn't even know that channel existed. Happy new year! [23:28] tomreyn: https://hastebin.com/apuxejucot.sql i find some uuid duplicate, but it was similar than another server that my friend have [23:28] *has [23:30] right, raid members in a mirror setup actually need to have the same uuid [23:31] *active