logs archiveIRC Archive / Freenode / #centos / 2015 / July / 21 / 1
jkyle
might be a red herring, that error, let me look
TrevorH
I don't think it's a gcc flag, that's why it throws it out
jkyle
yeah, reading, I don't think it's a failing error though. it just ignores the flag
but looks like someone's been....mucking with this system
ok, yeah, it's the flag causing the error
TrevorH
let me guess, someone's added export CFLAGS= to something in /etc/profile.d or in some start up script
jkyle
perhaps, #python's saying the flag was specified when python was built. I'm continuing my investigation
but, you certainly hit on something
yep, cflags set. fixed
I think, building
tuttles
I am seeing something strange in my Cent7 kickstarts. In my %pre section i have some tooling around destroying all the logical drives on the raid array and re-creating a new logical drive. Not every time, but most times when this in done in the %pre section the new logical drive shows up as /dev/sdb instead of /dev/sda, which causes all my subsequent kickstart commands to fail as they are coded around /dev/sda. Has anyone else seen this or worked around it?
metaspike
not sure, maybe you can use uuid's instead
they never change
tuttles
yea, i was thinking that, just more coding :-D
it did not happen in Centos6 so i think its more a kernel issue than an centos issue, but i am hopeful someone has seen this before
i see it on my dell and hp hardware
TrevorH
this is a hardware RAID setup?
         

hoherd
metaspike: i have not seen that problem, and i am using a modified centos 6 kickstart script for my centos 7 testing. i've done it with dell LSI raid and mdraid without problems.
oops, i guess that should've been directed at tuttles
TrevorH
and if you're booting off a usb stick or virtual media then that is probably sda
tuttles
hardware raid setup, it does not always happen which causes me to think its on my side, but its not consistent even on the same hardware
booting off pxe, so no usb or virtual media
at first i was only using dell so i thought it was a megacli problem with Dell LSI raid, but i just tried an HP G8 kickstart and got the same result
using Cent 7.1
theskillwithin
if you do mysqldump --user= --password= --databases db_name > skincare_live1.sql can you set -v so that it displays what its doing to the terminal, or will this cause the verbose to end up in the file?
TrevorH
tuttles: I'd add an fdisk -l or parted -l to my pre script and get the logs off it and see what _is_ sda
tuttles
yea, i have done that via a sleep 6000 and alt-f2 while kicking
its my disk
TrevorH
don't understand, I thought you said that was sdb at this point
tuttles
it starts at /dev/sda on boot, then in %pre i delete the logical drive and recreate it at the point of recreate it becomes /dev/sdb (i have a sleep 600 before i delete the logical drive, and a sleep 600 after it creates the logical drive)
list-hardrives even reflects the same thing
TrevorH
so what is sda at that point?
pj
I would put a sleep in between the delete and the create because it sounds like a race condition with dropping the old sda before creating the new one.
TrevorH
I have seen stuff like this with usb drives, plug it out/in and it gets the next drive letter up
tuttles
i do have a sleep between the delete and create (40 seconds)
sda is a gone after the create
TrevorH
then list-harddrives is your friend I guess
tuttles
i was thinking udev --settle or something
since its not consistent its no fun to trouble shoot, since hardware takes like 6 minutes to boot back up, its a little frustrating
pj
yeah, just gonna take time to troubleshoot, heh
TrevorH
where do you use the drive letter in your kickstart?
kexmex
can someone help in figuring out why cron stopped working?
pj
define "stopped working"
kexmex
pj: it's not firing anything hourly
i do "run-parts /etc/cron.hourly" manually, and it works fine
TrevorH
centos 6 or 7?
         

kexmex
i think i might have fixed it
CentOS7
i might not have had a line ending in 0hourly
i added it and i think that did it
TrevorH
that should be driven from /etc/cron.d/0hourly
kexmex
yea, in 0hourly
i didnt have a line ending after this line: 01 * * * * root run-parts /etc/cron.hourly
TrevorH
there isn't one in mine either but it still works
kexmex
TrevorH: i guess gonna see in an hour :)
TrevorH: i was deploying contents of that file via CHEF
TrevorH
not sure why you'd need to change it from the default out of the box that the package installs
kexmex
i added something
10 */3 * * * root run-parts /etc/cron.3hourly
is that not a good way to do it?
problem is 3hourly wasn't working inside cron.d
maybe it was missing that line ending too :)
TrevorH
why bodge it into an existing file when the whole idea of there being a cron.d directory is so that you can add your own files
kexmex
its true
if it works, i'll separate it
TrevorH
especially since that 0hourly file is not defined in the rpm as a config file so the next update to cronie will overwrite it
kexmex
why is it so damn finnicky
and no debug information
like crond would come up and not do anything at all-- it should have told me that it ignored that run_parts entry in there
TrevorH
pastebin that file
kexmex
well i already hit "enter" at the end of the line and saved it
TrevorH
also ls -la /etc/cron.d
do you know it's working ?
kexmex
http://pastebin.centos.org/31171/
-rw-r--r--. 1 root root 129 Jul 20 23:02 0hourly
TrevorH: if it's workign with the line ending? not sure yet
theskillwithin
on centos in my dev environment I want to create a script that turns on mysql logging the tail -f to the screen, then when im done it stops the mysql logging so I can quickly debugg problems
but I am afraid I will do more harm than good in trying to do this
kexmex
TrevorH: imma change time real quick and see if it runs
systemd: Reloading Command Scheduler.
lets see
scheduled for 26 min
you guys rewrote that stuff for cent?
hmm the box is showing 23:24 UTC
my lappie is showing 23:26
i guess i dont have ntpd?
this? ntp.x86_64
theskillwithin
yea nvm its too complex, I don't want to break something
kexmex
(root) CMD (run-parts /etc/cron.hourly) log_it: (root 57764) CMD (run-parts /etc/cron.hourly)
TrevorH: scripts in /etc/cron.hourly didnt run it seems
TrevorH
things to check: did you change the selinux context on 0hourly to something other than its default? are the files in /etc/cron.hourly executable? do they have the right selinux contexts?
kexmex
i think selinux is disabled
yea when i do run-parts /etc/cron.hourly manually it works fine
everything runs
how do i check selinux journal again- i forgot
TrevorH
getenforce
kexmex
Permissive
TrevorH
then it shouldn't be selinux
kexmex
yeh
it was working but stopped after "service crond restart "
« prev 1 2 3 4 5 6 7 next »