[00:01] never tried making something connect that way before. [00:01] my terminal knowledge is murky on that [00:02] wep or wpa? [00:02] wpa [00:02] static or dhcp address? [00:05] resno: http://pastebin.com/m4dec0024 [00:05] wow, thanks! [00:05] np...=) [00:06] didnt expect that [00:06] uh..forgot to translate: Mettere un IP coerente con la vostra rete -> Insert an ip fitting your needs..=) [00:07] what language is that? [00:07] italian [00:10] is libnss-ldap broken in 9.10? I've installed and configured, when I run a getent passwd I don't see the machine even attempting to make an ldap connection to get the information. What could I be doing wrong? [00:11] lol..i don't think they distribute unstable or broken packages, btw never used ldap or 9.10... [00:11] (9.10 cause of my fucking vps hosting company...) -.- [00:12] btw, going to bed...1:13 am here...cya tomorrow guys.. === alex88 is now known as alex88[sleep] [00:13] well I'm telling you it doesn't work. All my other Linux machines have no problems, yet the Ubuntu 9.10 machines can't do jack. [00:13] arctic_import: i've said i don't think..=) not it's impossible...good luck so.. [00:29] i've got a bunch of servers load balancing a website... is there a tool to consolidate log files from these various machines into one file/location? i was planning on rsyncing the files to an nfs share [00:51] _{pting: rsyslog and various other remote syslogging packages
[00:51] _{there's also splunk
[00:52] sub, thx, i'll check them out
=== matt_ is now known as Guest18838
=== Guest18838 is now known as aoluser24
=== dendrobates is now known as dendro-afk
[02:00] when running rsync as a daemon, does it sync up the files realtime?
[02:04] anyone have experience in net2ftp & ehcp?
[02:05] nvm... i'll just stick with non-daemon rsync cronjobs
=== erichammond1 is now known as erichammond
[02:41] New bug: #497087 in eucalyptus "Redundancy for cloud and cluster controller" [Wishlist,Confirmed] https://launchpad.net/bugs/497087
=== nxvl_ is now known as nxvl
=== jason__ is now known as neonfroen
=== neonfroen is now known as neonfreon
[03:21] Is BIND necessary if your domain registrar provides Full DNS control? Can't I just create A records on my Registrar and point them to the correct IP and call it aday?
[03:23] ChrisRut: Yes.
[03:24] ScottK: thanks
[03:24] BTW, how much resources does BIND use?
[03:24] * ChrisRut woundering if this will help save some load on my VPS (webserver)
[03:25] Hey how can I find network manager for pptp after install? It is not in menus?
[03:26] ChrisRut: It depends a lot on the query load. DNS tends to be more I/O bound than CPU bound.
[03:27] fbdystang: try installing network-manager-pptp
[03:27] fbdystang: sudo apt-get install network-manager-pptp
[03:28] ChrisRut: I did, but now i can't find the program to execute it
[03:28] ScottK: k, Thanks... Also what else does BIND do? Will I need to setup my MX records in my registrar aswell?
[03:29] ChrisRut: Yes. Any DNS records you publish. You'll also need to make sure your registrar points to their DNS servers as authoritative for your domain
[03:29] fbdystang: https://wiki.ubuntu.com/VPN
[03:30] ScottK: hmm, beginning to think this might be more trouble then it's worth.
[03:30] That's generally my view about bind.
[03:31] fbdystang: have you looked in network-manager for this? I am not sure, I've never used pptp
[03:32] ScottK: your view being? BIND=good or bad
[03:32] bind is good, but complicated.
[03:32] Generally, I suggest leaving running DNS to DNS experts.
[03:32] Hear, hear
[03:32] Your registrar may or may not qualify
[03:32] GoDaddy
[03:33] ChrisRut: It says that network manager is in the system tray, However i cannot find it. What does it look like? all my icons are accounted for
[03:33] There are two kinds of DNS, too -- publishing your domains' records, and caching everyone else's records.
[03:33] For the latter, dnsmasq is super nice.
[03:33] fbdystang: System>Preferances>Network Connections
[03:34] twb: but strictly for a webserver (ubuntu-server) why would I need to cache everyone elses?
[03:34] no outbound connections really
[03:34] ChrisRut: ok I found it, I see a vpn tab, but nothing on pptp
[03:35] webserver logging can actually do quite a lot of reverse dns lookups
[03:35] so sometimes caching of that is required
[03:35] or useful anyway
[03:35] ChrisRut: right; the latter only applies when you run a local network.
[03:35] really high-performance web serving generally turns off reverse dns lookups in the logs, but that's a bit annoying for normal production use
[03:35] ChrisRut: sorry, I wasn't paying much attention.
[03:35] fbdystang: try this:
[03:35] http://bit.ly/5RRFXU
[03:36] cjwatson: didn't think of that... true, thx
[03:36] cjwatson: if you're only caching for one host, nscd or whatever should handle it
[03:36] sure, I didn't specify a caching *method* :)
[03:37] ChrisRut just asked why caching would be useful
[03:37] OK
[03:37] ChrisRut: thanks, however i have already tried that one, to no avail :( I would think there would be a simple pptp gui to connect
[03:38] fbdystang: did you make sure to install 'network-manager-vpnc network-manager-openvpn
[03:38] I could be wrong, but i believe those packages add the ability to use VPN (therefore are required to use PPTP)
[03:39] Yea, thats what I would think too. But they are both installed
[03:40] is there a command line way to connect?
[03:41] fbdystang: how bout this one: http://bit.ly/6tvTkC
[03:42] hmm, thats new lemme check
[03:43] fbdystang: also give this a shot if that one doesn't work out: http://bit.ly/7rWItY
[03:47] OK, my network manager now has an icon in panel. I have created the vpn, but don't know how to connect it. it says last used is "never" wish there was a connect button!
[03:48] ChrisRut: will that second one work in karmic?
[03:48] 1sec..
[03:49] did you sudo /etc/init.d/networking restart
[03:49] after you setup the vpn
[03:49] go into terminal and issue 'sudo /etc/init.d/networking restart'
[03:49] It's probably going to kick you off irc thou
[03:49] I did sudo service network-manager restart, which it said is a more prefered one
[03:50] ok
[03:50] didn't kick me off though
[03:52] fbdystang: can you click on the network manager applet (top right) and then select your vpn connection that you just setup
[03:52] yes
[03:52] did that work?
[03:52] but i don't know how to connect it
[03:53] double clicking just brings up th edit window, which is where i set up the vpn
[03:54] How does it work in yours? does it just automatically connect or do you manually hit connect somewhere?
[03:54] not using vpn
=== dendro-afk is now known as dendrobates
[03:54] so I couldn't tell ya
[03:55] sorry, but I am stumped
[03:56] hmm
[03:59] in the second one it says to add my self to group 'dip' how do i do that on command line?
[04:00] sudo adduser $USER dip
[04:01] k thanks
[04:02] cjwatson: i'm getting a lintian error on eucalyptus-sc.templates: E: eucalyptus-sc: no-debconf-config
[04:02] cjwatson: can you confirm that this is okay?
[04:03] http://lintian.debian.org/tags/no-debconf-config.html
[04:03] cjwatson: says that it's sometimes okay, sometimes an error
[04:13] kirkland: it's fine, there's no db_input
[04:15] cjwatson: k, thanks; i'll add an override
[04:22] How do I know if the vpn connection worked?
[04:33] fbdystang: try pinging one of the boxes on your VPN
[04:40] gotta run, thanks for the help cjwatson, twb & ScottK
=== twb` is now known as twb
[06:58] I have a jaunty server; "sudo do-release-upgrade" says "After your package information was updated the essential package 'ubuntu-minimal' can not be found anymore." There are LP bugs about this, but they suggest it's been fixed. What can I do to diagnose?
[08:03] Mogge
[08:15] <_ruben> Mornin' Jeeves_
[08:48] morning folks
[09:15] <_ruben> bleh .. snort's an odd beast when you start playing with it
[09:18] :)
[09:44] Has any configured "dead peer detection" for their IPSec VPN on ubuntu server? I need to know if it is possible with racoon or do I have to use another package
[09:50] <_ruben> never used racoon (well, i did for a very brief period and hated it) .. openswan does have support for dpd
[10:04] _ruben when a package like openswan has dpd support, on being able to detect the dead peer is it possible to switchover to a secondary remote peer? What other corrective action can be taken. Thank you for your reply
=== _ is now known as Guest33639
[10:20] <_ruben> dpreacher: most common actions are "try to bring up tunnel again" and "put things on hold till its available again" .. openswan doesnt have high availability (switchover) builtin
=== alex88[sleep] is now known as alex88
[10:35] _ruben, can openswan call any other program to do switchover externally...like run a command
[10:39] <_ruben> dpreacher: there's updown scripts, but im not sure if those get called during dpd as well
[10:39] <_ruben> gotta go .. meeting
[10:39] seeya
[10:39] thanks
[11:39] are there any cases where "uptime" will report 2 users when there is only one logged in?
[11:40] 2 different users or 2x the same user?
[11:40] (use 'users' to check that)
[11:40] http://pastebin.com/m1e3868f8
[11:41] http://pastebin.com/m30c28257 <- updated with 'users'
[11:42] well, looks like only one user was logged in the second time
[11:43] anyway, "2 users" can mean "2x the same user"
[11:45] e.g. on my desktop I currently 3x my user logged in, and every terminal you start (e.g. in screen) adds another login
[12:13] JanC: no. the user "benedikt" is logged in once.
[12:40] morning
[12:40] kirkland, alive and kicking already?
[12:40] jiboumans: about to go for a run
[12:40] kirkland: ok, catch you after
[12:40] jiboumans: trying to make it back in time for the meeting
[12:40] morning zul
[13:43] Since an upgrade to karmic, ssh login to my server is quite slow. Any ideas?
[13:44] mathiaz: so what do I need for the scribe stuff?
=== zul_ is now known as zul
[13:52] kirkland: also it was supposed to require woodstox, but apparently doesn't need it... I guess that change hasn't landed yet in their code.
[13:52] all hail netsplits
[13:52] adac: does the login go as fast as usual (/etc/motd shows up etc, etc) but bash takes a second to load?
[13:52] benedikt, the moment I enter the ssh password in the console it takes up to 10 seconds until the "remote" bash appears
[13:52] 'uptime' is showing one user too many on my system ( http://pastebin.com/m30c28257 ) . What can cause this?
[13:53] benedikt, On jaunty it was there immediately
[13:53] adac: and th motd
[13:53] ?
[13:53] benedikt, what is that?
[13:53] adac: cat /etc/motd
[13:54] oh I see. Yes it takes 10 seconds until motd appears on login
[13:54] ah. then its not the same that started to happen to me.
[13:57] <_ruben> slow logins often tend to be dns related
[13:58] also, my issue about one login too many is resolved. i didnt realize that screen of course took up a tty and is in /var/run/utmp .
[13:58] zul: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase
[13:59] zul: outlines what to do as a scribe -basically write up the minutes
[13:59] mathiaz: k thanks
[14:15] in mail to mathiaz i suggested yaml as markup language, but python-yaml not in main. so thats a big strike against that.
[14:16] smoser: well - we may have to move to main anyway
[14:17] smoser: yaml is getting more and more popular
[14:17] smoser: I had a look at it this morning
[14:18] oh? in that case i'd consider, but the niceness of the built into python is nice. (ie, python-configobj)
[14:27] python-config is already in main as well
[14:29] smoser: python-yaml will probably come into Main due to a spec of mine. Also I'm the Debian maintainer for it, so I think it'll be easy to get into Main
[14:30] i really dont like the shortcoming of "some things can't be represented as keys" in python-configobj
[14:30] http://www.voidspace.org.uk/python/configobj.html#the-config-file-format
[14:30] There is no way of safely quoting values that contain both types of triple quotes.
[14:30] but maybe thats just worrying too much
[14:33] smoser: I think we should just try one
[14:33] smoser: once we have more experience we can add more syntaxes if needed
[14:33] agreed.
[14:34] did my mail to you make sense mathiaz ?
[14:34] smoser: yes
[14:34] smoser: I'll start by providing sample user-data script
[14:34] smoser: to see how things would work with the syntax (which ever syntax we choose)
[14:35] kirkland: hey dude, someone on the condor-users mailing list is asking about running condor on a eucalyptus cloud. thought you might like to know
[14:35] smoser: once we have the syntax and samples, we'll have tests as well and can then start implementing it
[14:35] mathiaz, thats fine.
[14:37] cr3: cool
=== SockPant1 is now known as SockPants
=== orudie_ is now known as oru_work
[15:03] anyone familiar with vsftpd? Wed Dec 16 09:51:27 2009 [pid 9660] [jpipublic] FAIL LOGIN: Client "75.127.142.196"
[15:04] same error on both the client and server log side
[15:05] soren, ping. could you sponsor bug 494185 to karmic-proposed?
[15:05] Launchpad bug 494185 in ec2-init "ec2-init selects us-east-1 mirror when running in us-west-1 region" [High,In progress] https://launchpad.net/bugs/494185
[15:07] smoser: i think its eod for soren as well
[15:07] so I could do that if you want
[15:07] well, then, do you want to do it? i was just hoping to spread the wealth
[15:07] sure
[15:08] dont you mean spread the pain ;)
[15:24] smoser: i have upload the ec2-init SRU i had to modify the debian changelog for it to be proper for an SRU
[15:25] what did you have to change?
[15:26] zul,
[15:27] i attached to the debdiff to the bug
[15:39] <`jpg> Hey guys, is it possible to install 9.10 on a machine without internet access?
[15:39] zul, so sru you're not supposed to increase ubuntu portion of the package version, but rather add '.X'?
[15:39] ie, you did: 0ubuntu7.1 rather than 0ubuntu8
[15:40] smoser: https://wiki.ubuntu.com/SecurityTeam/UpdatePreparation#Update%20the%20packaging
[15:40] smoser: ^^ has a table with expected version number for security updates
[15:40] smoser: SRU follow the same rules
[15:40] thanks mathiaz.
[15:40] smoser,: sorrect
[15:41] mathiaz: about the smart SRU it was done but it got rejected because there was no lp bug number in the changelog. I just re-uploaded it with the correct lp bug number
[15:42] zul: ok - can you keep an eye on it?
[15:42] mathiaz: will do
=== dendrobates is now known as dendro-afk
[15:50] hey folks. I'm porting a lot of code from windows to ubuntu and have a decent amount of unc paths being used for file access. In looking through the samba docs it looks like I'm going to have to create a mount point on my system for every other machine on the network i need to access... is that correct? Or is there some way I can just hit the shares via unc path name?
=== Guest13057 is now known as fahadsadah
=== dendro-afk is now known as dendrobates
=== fahadsadah is now known as Guest23237
=== Guest33639 is now known as viezerd
=== Guest23237 is now known as fahadsadah
[17:05] hi all, I am new to ubuntu UEC. I am trying to get elasticfox to manage the cluster. I am not able to add my cluster as a region in UEC. when i try accessing the URL "http://my_cluster_ip:8773/services/Eucalyptus" directly, it shows "Failure: 500 Internal Server Error". i tried elasticfox on both firefox 3.5 and then on 3.0, but to no avail. can anyone please help..
=== [1]pea_brain is now known as pea_brain
[17:10] hi all, I am new to ubuntu UEC. I am trying to get elasticfox to manage the cluster. I am not able to add my cluster as a region in _elasticfox_. when i try accessing the URL "http://my_cluster_ip:8773/services/Eucalyptus" directly, it shows "Failure: 500 Internal Server Error". i tried elasticfox on both firefox 3.5 and then on 3.0, but to no avail. can anyone please help..
[17:11] Is ubuntu server a decent server or should I go with something more along the lines of centos
[17:13] hiatus: ubuntu server ( LTS ) has 5 yrs support and centos has 7 yrs. besides, ubuntu has a lot of packages pacakged for it ( like 28000) while centos has about 3500 . just a couple of ideas from my side.
[17:13] Linux is Linux, whichever way you may wanna use it.
[17:19] hmmm, thanks pea, good points
[17:25] hiatus: one thing to note though, not all the packages i mentioned as supported on ubuntu have the same support commitments from the developers.
[17:28] how do you mean?
[17:34] there is a section called "main", which is fully supported. Others like "universe" and "multiverse" are not supported with the same level of commitment. http://www.ubuntu.com/community/ubuntustory/components
[17:55] kirkland: is there any WSDL specification for Eucalyptus?
[17:56] TeTeT_: i don't think i know enough about WSDL to answer that question
[17:56] kirkland: who could know about that, ttx maybe?
[17:57] TeTeT_: yeah
[17:57] TeTeT_: or ask in #eucalyptus
[17:57] TeTeT_: between network splits
[17:59] kirkland: I haven't thought of #eucalyptus :)
=== jcastro_ is now known as jcastro
[18:15] smoser: hi
[18:15] hey
[18:15] smoser: can I use the UEC images directly with kvm/libvirt?
[18:15] sort of.
[18:16] its a goal to answer 'yes, and heres how'
[18:16] smoser: ok - but we're not there yet
[18:17] you can do it, but ec2 init is dependent upon the metadata service
[18:17] so you either have to
[18:17] a.) give a fake metadata service
[18:17] b.) disable ec2-init
[18:17] c.) wait 30+ minutes for ec2-init to time out
[18:17] https://wiki.ubuntu.com/UEC/Images/Testing is mostly still relevant i think
[18:18] the piece of info missing there is that you can append to the kernel 'ec2init=0' and that will disable ec2-init
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== astechgeek is now known as techgeek
[18:27] New bug: #497495 in eucalyptus "euca-delete-bundle does not correctly parse dir path in some cases" [Low,Fix committed] https://launchpad.net/bugs/497495
[18:56] can someone recommend an intranet solution ?
[18:57] errr... you'll have to be a *lot* more specific than that. ;)
[19:13] hi - trying to setup a LAMP server using Ubuntu server ... anybody had any problems using Ubuntu 9.1 server in fusion
[19:13] the server seems to be for AMD 64 bit
[19:17] you mean 9.10 right?
[19:17] and there should be a 32-bit server ISO download.
[19:18] 64 bit is the default, because generally there aren't any 32-bit only servers anymore. ;)
[19:18] yes 9.10
[19:18] trying to find 32bit
[19:18] just trying to test something at home u c
[19:19] was searching on the web site and came across this channel - so thought I'll ask
[19:20] but all the links takes me back to the 64bit d/l .... ummmm
[19:21] let me find you a link.
[19:21] thanks :-)
[19:21] http://www.ubuntu.com/getubuntu/download-server
[19:21] you see under the download button, it says "Alternative download options"?
[19:22] Click that, then you can select the 32 bit version.
[19:22] oh thanks - so stupid ... should of saw that
[19:22] no prob. :)
[19:22] also, if you go to one of the mirror sites directly, there'll be bittorrent download links as well, if that's faster for you.
[19:23] torrent no good for me as my ISP throttles those type of traffic
[19:23] but thanks for the suggestion
[19:23] yeah, some ISPs BT is way faster, some it's better to just use the local mirror. the options are there, anyway.
[19:24] :-)
=== rickspencer3_ is now known as rickspencer3
=== erichammond1 is now known as erichammond
=== NCommander is now known as Guest36380
=== Guest36380 is now known as NCommander
=== erichammond1 is now known as erichammond
=== erichammond1 is now known as erichammond
=== jussio1 is now known as jussi01
=== erichammond1 is now known as erichammond
=== erichammond1 is now known as erichammond
[21:02] kirkland, have you found a way to get -curses to work with lucid ?
=== ikonia_ is now known as ikonia
[21:28] Has anyone arranged for ZFS in Linux allready? :)
[21:28] (not using fuse)
[21:30] sorry to tell you, it won't happen until linux changes it's license, gpl gets reworded or new version, or zfs's license changes
[21:30] and none of those will be happening any time soon
[21:34] c'è qualche italiano collegato?
[21:47] smoser: unfortunately not yet
[21:54] yea ubuntu server room
[21:54] my favorite of all
[22:00] :)
[22:00] :)
[22:00] i got 8.04 server running atm
[22:01] epinky: you run server
[22:02] anyone else her
[22:02] here**
[22:03] My headless server keeps shutting down - what can I do to figure out why it is shutting down on its own?
[22:03] bad memory
[22:03] is a possiblity
[22:03] for me it is cause 1000+ memtest errors
[22:04] but it runs just when i restart from time to time it doesn't start up again until i unplug it and plug it back in
[22:04] khelvan have you looked through the logs?
[22:04] that's one way to find out
[22:04] didn't think bout the logs
[22:04] dunix - I am not familiar with where to find logs or what to look for. I am brand new to this.
[22:05] overheating is another common reason
[22:05] and bad mem
[22:05] you come a across bad address and bam down
[22:05] /var/log/messages is a good place to start
[22:05] or /var/log/syslog
[22:05] well i shall revise that
[22:05] bad addresses and sometimes bam
[22:11] Here is the /var/log/messages - I don't see anything except a random restart, which I have no idea why it did: http://pastebin.com/m3f44bf33
[22:12] do you have it setup to automatically restart
[22:12] generally people set it to do this to avoid errors
[22:12] much like with cisco routers
[22:13] Here is the syslog, with another random restart - http://pastebin.com/d37f72e41
[22:13] No, I never set it to do that, it has been running fine for a month, and now all of a sudden it is randomly restarting (and randomly shutting down, as in power off).
[22:14] did you do a memtest
[22:14] just to make sure
[22:14] can never be to careful
[22:14] I have not, no, how do I do that with a headless server?
[22:15] well if you can restart the machine with a monitor attached you can press escape key when it says please wait 3 seconds at the grub boot loader menu
[22:15] then boot to memtest
[22:15] my server restarts from time to time but thats cause my mem is finished
[22:21] khelvan, is this hardware built to be a server?
[22:21] ie: a dell or hp server etc, or is it a desktop you are using as a server?
[22:21] dunix - It is desktop components I put together to be a server.
[22:22] you may want to check your bios to see if there is any temp settings, usually with a problem like this, it's a heat issue
[22:23] and some bios' have settings for temp, even the desktop ones
[22:23] or you may just want to monitor the internal temp of your MoBo/Proc, if it allows for that
[22:23] dunix - Are there any temperature monitors I can install and watch via ssh?
[22:23] lm-sensors
[22:24] but may or may not work, depending on your motherboard
[22:24] Ok, it is already installed - how can I monitor it via SSH?
[22:24] Google+lm-sensors website would tell you if your stuff is supported
[22:25] i was talking monitor like a crt or lcd
[22:25] but ssh
[22:26] Anonymous2 - I know, it isn't practicle for me to have a physical monitor on the machine at the moment. I am installing/monitoring/checking the logs via ssh at the moment.
[22:26] (practical)
[22:26] you can't monitor the memtest
[22:26] you can with a computer monitor like lcd screen and keyboard
[22:26] it's hard for me to explain
[22:27] I realize that, it will have to be a bit before I can run that. I understand, I have run memtest before, which is why I asked how I can do it with a headless server.
[22:27] The short answer is that I can't.
[22:27] khelvan, if everything is working properly i believe you can type 'sensors'
[22:27] It needs a head.
[22:28] dunix - Thanks, I got this output: k8temp-pci-00c3 Adapter: PCI adapter Core0 Temp: +33.0°C
[22:29] Is there a way for me to tell if it was a system-initiated clean shutdown, or if it crashed due to heat/memory? I have some critical data on a RAID array, how can I insure that the file system is not going to be corrupted?
[22:29] well thats not anything to worry about, is there any way/reason this would spike? like for example a dog sitting behind it blocking the intake (which is what my dog does)
[22:29] if was a system, it would have said in the logs
[22:30] you can check on the RAID with /proc/mdstat
[22:30] it'll automatically rebuild if something is wrong
[22:30] as for how dangerous it is, that depends on what filesystem you use
[22:31] ext3 is generally fine with sudden failures
[22:31] might lose some data from writes occurring during the failure
[22:31] but otherwise, I've never lost anything else on ext3
[22:31] XFS is the opposite, I've had several catastrophic failures in crashes
[22:33] dunix - No pets, the PC is sitting on a stone floor in about 69-70 F heat
[22:33] Or temp, I should say
[22:33] qman__, it is ext3
[22:33] then i'd say temp is ruled out
[22:34] qman__, how do I check with /proc/mdstat?
[22:34] cat /proc/mdstat
[22:34] that's only if you're using mdadm/linux soft raid
[22:34] if you've got a hardware raid or a fakeraid that won't tell you anything
[22:35] Err, I am using the RAID that I automatically set up during server install at the prompts, which I believe is linux soft raid
[22:35] yeah
[22:35] that's mdadm
[22:35] qman__, output: md0 : active raid1 sdc1[0] sdd1[1] 488383936 blocks [2/2] [UU]
[22:35] that means it's fine and online
[22:35] it'll say rebuilding if it is, and it'll show one of the disks failing if it is
[22:36] I guess I better get that backup solution up and running soon just to be sure.
[22:37] backups are always a good idea ☺
[22:38] this is what my file server looks like
[22:38] md0 : active raid6 sdh1[7](S) sda1[0] sdg1[6] sdf1[5] sde1[4] sdd1[3] sdc1[2] sdb1[1] 4883799680 blocks level 6, 64k chunk, algorithm 2 [7/7] [UUUUUUU]
[22:38] JanC - Yeah, I agree, it has just been low priority for me with all sorts of other problems going on at the moment with other stuff.
[22:39] qman__, What does the (S) refer to?
[22:39] spare
[22:39] previously, I had just raid 5
[22:39] and lost two disks within an hour
[22:39] suffered total data loss
[22:39] so, this time, raid 6 with a spare
[22:40] 2 disks from the same series?
[22:40] same model, bought at the same time
[22:41] ran for about two years
[22:41] it's quite common that they break around the same time then
[22:41] Wow, aren't the odds of that astronomically high?
[22:41] then died, an hour apart
[22:41] qman__: I once heard somebody tell they lost 3 disks within 5 hours that way ;)
[22:42] was an ISP's mailserver...
[22:42] ouch
[22:42] lol
[22:42] So the moral of the story is buy two different types of disks?
[22:42] luckily my data wasn't that valuable
[22:42] the moral is, keep backups
[22:42] all disks die eventually
[22:42] what qman said.
[22:42] well, the problem with different disks is that they have different timings
[22:43] yeah, the array will perform better if you use the same disks
[22:43] _{and different geometries as well, yeah?
[22:43] AFAIK "geometries" are all virtual these days
[22:44] this time around I've got it set up to email me on my phone if it detects a SMART error, so that shouldn't happen again
[22:44] unless of course you have a support contract, then you have to use identical drives
[22:44] qman__: that doesn't always work
[22:44] SMART can't predict all errors
[22:45] yeah
[22:45] mdadm is also set to email me on failure
[22:45] but SMART should pick it up first
[22:45] and raid 6 with the spare should buy me more time than an hour
[22:45] I had one disk where SMART gave no warning until the disk was dead
[22:46] disk was 1 month old...
[22:46] there's always a chance for total loss, but I'm trying to improve the odds
[22:47] the 4GB Quantum Fireball I'm using as a system drive out lasted the first array
[22:48] if/when that one finally dies I'm switching to flash
[22:48] but all signs point to that thing lasting longer than me}}}