neonixcoder | I am facing an issue with my cmos resetting to 2009-01-01 time as mention hear http://www.vistax64.com/general-discussion/286845-help-cmos-resetting-problem.html | 00:16 |
---|---|---|
neonixcoder | I checked my CMOS battery and its voltage shows 3.3v. | 00:17 |
neonixcoder | I dont see any issue with CMOS battery.. any other suggestions? | 00:17 |
sarnold | measuring the voltage of a battery without a load usually doesn't work | 00:19 |
sarnold | you might as well just replace the thing outright as a first troubleshooting step, they're only a few bucks | 00:19 |
neonixcoder | sarnold: I checked with load and without load.. they are much identical.. | 00:22 |
neonixcoder | Yes.. but before replacing I want to conform if that is a CMOS battery issue.. | 00:23 |
neonixcoder | I have total 200 ubuntu boxes which are running and I see this issue on 20 boxes | 00:24 |
sarnold | were they all bought at the same time? :) | 00:24 |
sarnold | it really could be a software issue, if something is setting the time on your running systems to six years ago, and then running hwclock --set | 00:25 |
neonixcoder | At different times, different times deployments and present in different locations. but with same configuration. | 00:25 |
sarnold | but I think dual-boot systems have severe trouble with hwclock --set, so I don't think it's acutally used by default any more | 00:25 |
neonixcoder | sarnold: If I am not wrong, ntp will set hardware clock right? | 00:26 |
patdk-lap | if you never set the clock, and hever use hwclock --set, could be an issue, but it should never reset to 01-01 again though | 00:26 |
sarnold | neonixcoder: I don't think so, I think it only sets the software clock | 00:26 |
patdk-lap | it would atleast still be counting up | 00:26 |
patdk-lap | so the issue is with the rtc or the battery | 00:26 |
patdk-lap | it could be as simple as someone has the jumper on the clear cmos, still | 00:26 |
sarnold | neonixcoder: replace two or three of the twenty, see what happens; I bet those are fixed. | 00:27 |
sarnold | patdk-lap: hah :) | 00:27 |
patdk-lap | on my newer boards, it's just two solder pads | 00:27 |
patdk-lap | might be a solder bridge on them | 00:27 |
sarnold | no pins? | 00:27 |
patdk-lap | no pins | 00:27 |
patdk-lap | you just touch it with a screwdriver or something | 00:28 |
patdk-lap | I guess that was to keep people from leaving the jumper on | 00:28 |
neonixcoder | sarnold: sure will check on couple of machines before replacing all | 00:28 |
=== markthomas is now known as markthomas|away | ||
=== cipi is now known as CiPi | ||
lordievader | Good morning. | 08:01 |
patsToms | lordievader, morning | 08:04 |
lordievader | Hey patsToms | 08:06 |
=== Lcawte|Away is now known as Lcawte | ||
=== pv2b_ is now known as pv2b | ||
=== kickinz1 is now known as kickinz1|lunch | ||
rbasak | hallyn: could you take a look at bug 1483836 please? It's in main (ubuntu-server subscription) because it's a build-dep for qemu AFAICT. | 11:05 |
ubottu | bug 1483836 in acpica-unix (Ubuntu) "acpica-unix ftbfs on powerpc" [High,Confirmed] https://launchpad.net/bugs/1483836 | 11:05 |
rbasak | hallyn: I can help you with the powerpc porter if you need it. | 11:05 |
=== Guest55878 is now known as IdleOne | ||
=== kickinz1|lunch is now known as kickinz1 | ||
hallyn | rbasak: after libvirt is straightened out | 13:38 |
rbasak | hallyn: np, thanks | 14:05 |
hallyn | zul: smb: if you have a minute, could you look over http://paste.ubuntu.com/12061910/ ? Final patch which I intend to push to libvirt this morning | 14:09 |
zul | hallyn, sure | 14:09 |
zul | hallyn: looks ok | 14:11 |
hallyn | ok thx | 14:12 |
smb | hallyn, Hm yeah nothing that immediately jumps out. Though large enough to be hard to say completely for sure | 14:18 |
badabum18 | Hi, is there anybody to tell me with anacron 2.3 variable RANDOM_DELAY ? My job appears as "Will run job `test' in 1 min." in syslog but 1 mins is the period and RANDOM_DELAY=120 is ignored... :-/ | 15:13 |
Overand | Should I stick my SSL settigns (ala disabling SSLv3) in my /etc/apache2/mods-__/mod_ssl.conf - or in my virtualhost? | 15:24 |
Overand | seems like it should go in the module settings, but is that going to get axed by package management? | 15:24 |
=== esde is now known as esde1 | ||
jelly | Overand: put it in the ssl virtualhost | 15:49 |
=== markthomas|away is now known as markthomas | ||
Dr_Apocalypse | Greetings all. I'm wondering if anyone has some suggections for setting up a reverse proxy | 16:43 |
pmatulis | Dr_Apocalypse: study nginx | 17:36 |
=== esde1 is now known as esde | ||
Qantourisc | Since I installed ubuntu i used bootstrap on a dir. So i'm not sure everything was installed correctly. | 18:17 |
Qantourisc | What is the default syslog used by ubuntu ? | 18:18 |
Qantourisc | ok enoughg google results seem to point to rsyslogd | 18:19 |
tarpman | Qantourisc: install the package "ubuntu-minimal" to get the basic system packages such as syslog and cron | 18:24 |
Qantourisc | tarpman: well, currently that would take ... I don't know ... 3 hours, so first installing some debug tools :p | 18:24 |
Qantourisc | but thanks | 18:24 |
Qantourisc | lxc is acting up weird somehow | 18:25 |
Qantourisc | aptitude/apt-get is verry slow, diks usage, it claims | 18:25 |
RoyK | Qantourisc: are you low on diskspace? | 18:26 |
Qantourisc | no but I did just add storage | 18:27 |
RoyK | things are usually rather slow if filesystems are almost full - leads to very fragmented filsystems (on all OSes) | 18:27 |
Qantourisc | also tty1 refuses root login | 18:27 |
Qantourisc | RoyK: not talking slow, talking floppy disk slow | 18:27 |
RoyK | how did you add storage? | 18:27 |
Qantourisc | lvresize; resize2fs | 18:28 |
RoyK | should work | 18:28 |
Qantourisc | it did :) | 18:28 |
RoyK | and should help writes do better, though reads may still be in trouble if the filesystem is highly fragmented | 18:29 |
Qantourisc | the FS is quite fresh, and never got near full | 18:29 |
RoyK | is this a vm or physical? | 18:29 |
Qantourisc | vm in lxc | 18:29 |
Qantourisc | take ~1 minute to "build dependency tree" | 18:30 |
RoyK | anything from dmesg about bad drives? | 18:30 |
Qantourisc | shouldn't be | 18:30 |
RoyK | well, check | 18:30 |
Qantourisc | just did: no | 18:31 |
Qantourisc | however it is loading 1 disk of the mirror raid more then the other | 18:31 |
Qantourisc | 75% load on 1 16% on the other | 18:31 |
Qantourisc | that is weird | 18:31 |
RoyK | pastebin /proc/mdstat | 18:31 |
Qantourisc | http://codepad.org/DjNkXNuV | 18:32 |
Qantourisc | nothing in it btw :p | 18:32 |
RoyK | 64MB chunks??? | 18:33 |
Qantourisc | it's reading @130MB's how can that even be | 18:33 |
RoyK | bitmap: 2/2 pages [8KB], 65536KB chunk | 18:33 |
* Qantourisc rechecks the chunk size, that does sound large :D | 18:33 | |
RoyK | seems to be the bitmap chunk | 18:34 |
Qantourisc | RoyK: try looking at md127 | 18:35 |
Qantourisc | i mean 124 | 18:35 |
Qantourisc | 7566523392 blocks super 1.2 level 5, 512k chunk, algorithm 2 [3/3] [UUU] | 18:35 |
Qantourisc | the chunksize there is 512k | 18:35 |
Qantourisc | the rest must indeed be related to bitmap | 18:35 |
RoyK | yeah, but the bitmap chunk size is 64MB | 18:35 |
RoyK | doesn't seem to be the problem, though, although you might want to ask on #linux-raid | 18:36 |
Qantourisc | I did a apt-get install rsyslog about 6 mins ago, it's at "Reading state information... Done" atm, to give you an idea how slow btw :) | 18:36 |
RoyK | better place to ask about raid stuff than in here | 18:36 |
Qantourisc | sec idea | 18:37 |
Qantourisc | where are the "data" files of apt located ? | 18:38 |
RoyK | iirc /var/cache/apt/ | 18:38 |
Qantourisc | ok looks right | 18:38 |
Qantourisc | just catted all that data to /dev/zero, verry fast :p | 18:39 |
* Qantourisc smells exesive IO due to something else failing | 18:39 | |
dasjoe | apt is calling a lot of fsync, so you might want to get libeatmydata and preload that ;) | 18:39 |
Qantourisc | dasjoe: hmm good idea to debug | 18:40 |
* Qantourisc tries | 18:40 | |
RoyK | using a mirror of an ssd and spinning rust for the root with write-mostly for the spinning rust works well | 18:40 |
Qantourisc | i'll have the results in ~10 mins | 18:40 |
Qantourisc | takes a while to install :/ | 18:40 |
Qantourisc | RoyK: ... it's on ssd ... (entirely) | 18:41 |
RoyK | and it's that slow?!? | 18:41 |
Qantourisc | yes | 18:41 |
Qantourisc | hence my supprise :) | 18:41 |
RoyK | better run a smartctl -H /dev/whatever | 18:41 |
RoyK | guess it may be bad | 18:41 |
dasjoe | "LD_PRELOAD=libeatmydata.so aptitude" | 18:42 |
Qantourisc | dasjoe: it's not installed yet | 18:42 |
Qantourisc | hmmm maybe i can install it from the host lxc server | 18:42 |
dasjoe | Once libeatmydata is installed. Also, I've got a tiny script for checking interesting SMART stats: https://gist.github.com/dasjoe/a7252e3f737ec36de36f | 18:42 |
Qantourisc | yea host system is chroozing by | 18:43 |
Qantourisc | copied eatmydata, no real improvement | 18:44 |
RoyK | Qantourisc: if you have i/o issues like that on an ssd, make sure you have a good backup | 18:45 |
Qantourisc | tty1 is also borked: when i type root, it rejects the user | 18:45 |
RoyK | Qantourisc: looks to me that ssd is borked | 18:45 |
dasjoe | Qantourisc: so, the host is fast but the container is slow? | 18:45 |
Qantourisc | dasjoe: es | 18:45 |
dasjoe | Qantourisc: if that's the case your SSD is fine but your LXC is weird. I have no experience with containers, so I'm not very helpful, sorry | 18:46 |
RoyK | Qantourisc: pastebin 'smartctl -x /dev/sdX' for that ssd | 18:46 |
Qantourisc | think i must have some missing resource/functionality on the container | 18:46 |
Qantourisc | RoyK: it's 2 :p | 18:46 |
RoyK | Qantourisc: those SSDs may sometimes not report things right | 18:46 |
RoyK | Qantourisc: a mirror? | 18:46 |
Qantourisc | yep | 18:46 |
dasjoe | RoyK: it's not related to hardware, imho. Their host is running fine, but IO from within the container is slow | 18:46 |
RoyK | well, that's good | 18:46 |
Qantourisc | RoyK: not really | 18:46 |
Qantourisc | it's the same brand, and order | 18:47 |
Qantourisc | that's bad | 18:47 |
dasjoe | Qantourisc: hop over to #zfsonlinux and let's get you up and running with a real file system ;) | 18:47 |
RoyK | Qantourisc: without write-mostly, all i/o, also reads, will be spread to all drives | 18:47 |
RoyK | dasjoe: linux filesystems also work - this thing seems to be hardware | 18:48 |
RoyK | dasjoe: also, I've been using zfs for 6+ years ;) | 18:48 |
Qantourisc | dasjoe: tried zfs once,has some concept issues with ot | 18:48 |
Qantourisc | RoyK: still want those smarts ? I uploaded them | 18:48 |
Qantourisc | http://codepad.org/dIK1h9da http://codepad.org/KHctwXIV | 18:48 |
RoyK | please | 18:48 |
Qantourisc | doubt you will find anything :p | 18:49 |
Qantourisc | they are warmer then expected though :p | 18:49 |
Qantourisc | then again the chasis is not closed :p | 18:49 |
Qantourisc | it's running at 2800 reads / second | 18:50 |
RoyK | Qantourisc: doesn't look like anything bad - but does iostat say anything about which ssd is spending more time? | 18:50 |
Qantourisc | readking 150 000 kB/s | 18:50 |
Qantourisc | /dev/sdd | 18:51 |
Qantourisc | but it's also getting the bulk of the read-requests | 18:51 |
RoyK | try to detach it or offline it | 18:51 |
Qantourisc | wich is verry weird | 18:51 |
RoyK | mdadm --fail iirc | 18:51 |
Qantourisc | Such a hassle to get it back online, how certain are you to persue the pure-hardware-io path atm ? | 18:52 |
dasjoe | Qantourisc: check the disks's readahead, I've seen IO thrashing due to linux assuming sequential reads when I had random I/O with large enough block sizes | 18:53 |
RoyK | just wonder how 150MB/s can't be sufficient | 18:53 |
Qantourisc | RoyK: I think it's stuck on something due to something msissing / me takes a look at strace again | 18:53 |
Qantourisc | are mmap -read mentioned in strace | 18:54 |
Qantourisc | i'm not entirely sure | 18:54 |
dasjoe | blockdev --getra /dev/sdX | 18:54 |
Qantourisc | dasjoe: you'd want to check /dev/sdX and the lvm btw | 18:54 |
Qantourisc | and getfra too iirc | 18:54 |
RoyK | Qantourisc: perhaps someone is there doing something they shouldn't? box rooted? | 18:55 |
Qantourisc | god lets hope not | 18:55 |
dasjoe | RoyK: I still think it's in the container layer, the host is fast on the same HW, if I understand correctly | 18:55 |
RoyK | Qantourisc: can you clarify that? | 18:56 |
Qantourisc | RoyK: well for starters, there is no direrct exposure to the internet | 18:56 |
RoyK | Qantourisc: is the host responsive? | 18:57 |
Qantourisc | they'd have to have access to another system in the network | 18:57 |
Qantourisc | RoyK: yes | 18:57 |
Qantourisc | IO == 0 when aptitude is not running | 18:57 |
nettlejam | Does anyone know why the Ubuntu EC2 AMI locator (http://cloud-images.ubuntu.com/locator/ec2/) lists 15.04 as 'DEVEL'? | 18:57 |
RoyK | Qantourisc: I don't use lxc so really, I don't know, but have you tried restarting the guest? | 18:59 |
Qantourisc | RoyK: relucantly yes, even unmounted and did a fschk | 18:59 |
linocisco | how to use aptonCD on ubuntu server? | 19:00 |
Qantourisc | Can I disable mmap in apt-get ? | 19:00 |
RoyK | you shouldn't | 19:00 |
RoyK | if i/o is the problem, iotop might show something | 19:01 |
Qantourisc | RoyK: well there are no reads when using strace, since it's mmap-ing, so that would allow me to see some reads | 19:01 |
linocisco | how to use aptonCD on ubuntu server? | 19:02 |
Qantourisc | yea defently lxc => chroot is fast | 19:06 |
linocisco | https://help.ubuntu.com/community/APTonCD is for Desktop, what about for Server version? | 19:09 |
patdk-wk | how did you define *desktop* | 19:10 |
patdk-wk | where does it say, for desktop only? | 19:10 |
sarnold | patdk-wk: he left.. he lacks patience. | 19:10 |
patdk-wk | oh, It was so quick, I didn't notice | 19:10 |
patdk-wk | damn it, now my inner troll is going build up | 19:10 |
sarnold | hehe I know the feeling... I was all excited to type out "what error messages did you get when you tried?" oh well | 19:11 |
=== [1]badabum18 is now known as badabum18 | ||
host127 | Anyone may explain to maiself if this: > http://store.bq.com/gl/smartphones have sudo ? | 19:37 |
sarnold | host127: yes, it does | 19:39 |
sarnold | host127: you may have better luck in #ubuntu-touch | 19:39 |
host127 | sarnold: don lie to me. please. because if is true i will buy ok? | 19:40 |
sarnold | host127: I just tried it on my nexus4 running ubuntu touch; when you start the terminal, it even prints 'To run a command as adm8nistrator (user "root"), use "sudo <command", See "man sudo_root" for details.' | 19:41 |
dasjoe | sarnold: how's the battery life? | 20:29 |
sarnold | dasjoe: I don't have a SIM card in that phone, so it's not really a fair comparison, it never does actual phone things | 20:31 |
sarnold | dasjoe: it runs several days between charges | 20:31 |
dasjoe | sarnold: oh, just like my OPO. I use it for reading tposana in Play Books ;) | 20:33 |
sarnold | dasjoe: my ubuntu phone has mostly turned into a machines vs machines game :) | 20:34 |
sarnold | it's yet another tower dfense game, but you've got to get tower placement, types, and upgrades done just right or you fail miserably | 20:34 |
dasjoe | Sounds interesting. I really fell for the Kingdom Rush series | 20:37 |
=== markthomas is now known as markthomas|away | ||
Qantourisc | RoyK: cause is some cgroups going haywire | 21:04 |
Qantourisc | RoyK: not sure why or wich one yet | 21:05 |
Qantourisc | RoyK: memory limit | 21:09 |
Qantourisc | RoyK: limit was 12MB | 21:11 |
=== markthomas|away is now known as markthomas | ||
Qantourisc | So to fullfill the mmap it has to swap in rather small segments of file | 21:13 |
Qantourisc | dasjoe: Ping read above if you care about what the issue was | 21:13 |
sarnold | _twelve megabyte_ cgroup memory limit? man, how'd you do that? :) | 21:17 |
sarnold | and how on earth did anything run? | 21:18 |
=== IdleOne- is now known as IdleOne | ||
a1fa | hello, i am trying to install ubuntu 14.04 on HP Proliant G8 with HP p420i RAID and SAS controller.. i removed my disks from raid configuration because i want to do btrfs, but the drives dont show up during install, and i am prompted for driver install because no disks have been found | 22:51 |
sarnold | a1fa: does your controller have an HBA mode or allow you to flash IT firmware? | 22:52 |
a1fa | i did not see that option | 22:53 |
a1fa | i'll check again | 22:53 |
a1fa | well looks like its going to be a command line thing using their cli tool | 22:54 |
=== Lcawte is now known as Lcawte|Away |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!