logs archiveIRC Archive / Freenode / #centos / 2015 / July / 2 / 1
kbabioch
hi ... i have an centos installation with "net.ipv6.conf.all.disable_ipv6 = 1" everytime i reboot. can't find where this setting is made, any hints?
Guest40597
kbabioch: /etc/sysctl.conf
kbabioch
wN: No, actually there was a sysctl file in the initramfs
I removed it on the file system already, but didn't rebuild the initramfs
that did cost me some time to figure out, though ;)
but thanks for your response
wN
kbabioch: the sysctl on the filesystem should overwrite the initramfs one, should it?
shouldnt it*
also, the sysctl in the initramfs is also named /etc/sysctl.conf
hehe
kbabioch
wN: No, there was a dedicated sysctl file called "01-disable-ipv6.conf", which was created manually before. now i wanted to re-enable ipv6, but simply removing the file didn't help, because it was still in the initramfs
wN
ah i see
OH
i gotcha
yeah, you shold have set it to 0 on the fs
that'd probably do it
kbabioch
yup, it would, but i don't like these kind of modifications, since "enabled" is the default ...
nobody will understand what i was trying to solve in a couple of months and even i won't remember :D
GlenK
hey, anyone know how to change the gdm background on centos 7? nothing I'm finding via google is working out for me.
closest I've come is this dconf update thingy after mucking about in /etc/dconf, but that just replaces the small centos logo at the bottom with my wallpaper
ofir
guys any idea why sudo reboot never actually reboots, and I have to cold reboot the machine to get it to actually _reboot_?
I'm guessing something is blocking the restart but not sure how to debug it
         

wolfy
ofir: any AVC denials ?
btw, reboot should work withiout sudo, it relies on consolehelper ("the userhelper program [...] used to allow configured programs to be run with superuser privileges by ordinary users.")
pj
ofir: try: sudo shutdown -r now
that may or may not work
at the end of the day what either does, I think, is to change the systemd target (or in older CentOS the runlevel)
parallel21
Does networkmanager have to be disabled for bridge ethernet to work?
fenrus02
depends what version of centos you have .. in 7, nm should be able to do a bridge. in 6, you'll need to tell your interface *not* to use nm
parallel21
Hrmm...
I was following these instructions
http://www.certdepot.net/rhel7-configure-lab-network-settings/
But I don't understand why they are doing so
fenrus02
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Networking_Guide/ch-Configure_Network_Bridging.html
parallel21
fenrus02: You rock, I suck at the google
fenrus02
google "rhel 7 bridge networkmanager" produced that link as #1
occupant
just turned off a fc6 server with 1900 days uptime
hooray for eventual responsibility
DiscordianUK
Long while without updates
fenrus02
barely under 8 yrs dead ..
could have done c5 instead, and had fixes aplenty with the same api as fc6 had
DiscordianUK
Indeed
occupant
nope, c5 came out a full year afterward and this install happened inbetween
c4 still would've been a better option, but I wasn't around for all this anyway
DiscordianUK
I became aware of centos at c5
hmm actually I mean c4 as that was what I deployed
occupant
I had to deal with an old c4 at my last job... on boot, it complained that more than 8GB of memory was an unsupported config. how quaint.
grobe0ba
anybody know anything about any (spoken all of them) out-of-sync mirrors? (CentOS 7.1 x86_64)
wait
nm. i just realized i did something stupid
pj
grobe0ba: epel is having issues.
grobe0ba
pj: nah, was my mistake
pj
ok
grobe0ba
happens when you don't assign a default route :P
         

pj
ahhhh, ok
well been getting a lot of complaints lately about yum blowing up due to the epel issues.
so it's easy to make the assumption.
grobe0ba
yeah
autofsckk
good night, i have centos 6 i686 installed on a vps, and i need to change it to version 7 64 bits, whats the best way to do it? how can i get a list of installed apps? should i just reinstall everything and put back /etc on place and thats it?
kaos01
does using hp virtual connect flexfabric fcoe turn up in the linux as a fceo0 interface ?
Motoko
I'd install from scratch.
Changing arch is going to be very tough, and the normal upgrade process isn't the easiest to begin with.
rpm -qa
That'll list all installed RPMs.
Also, don't overwrite /etc, EL7 changed a lot of things and you'll break things if you do that.
You'll need to just copy back the configs you need after verifying that they're compatible.
autofsckk
a lot of work
Motoko
Yes, it will be.
It might be best to test on a clone.
autofsckk
i just have a production linode small vps and cant afford to have another to test on, i think that ill just have to shut it down at night and do things fast ajjajajajaja :/
Motoko
Good luck.
autofsckk
is there a way to do it locally, i mean install everything and just upload config files and stuff to the vps?
is that waht you are talking about a clone?
Motoko
Well, you can test your procedure by using a local VM.
autofsckk
i wont be able to do it cause my laptop has a 32 bit cpu :/
pj
autofsckk: you'll be dead in the water if you try to upgrade a linode from el6 i686 to el7 x86_64 in place, I would say odds are you wont' succedd and will end up with a bricked linode.
autofsckk: if you can't afford a second linode then I would just continue to stick with el6 for a while.
Motoko
(Action) nods
el6 will be supported for a bit longer. If you need newer versions of web components, IUS is a nice repository.
pj
there are a few 3rd-party repos out there with newer stuff for el6
autofsckk
pj: its more important for me to change from 6 to 7, is that possible? secure? the change from 32 to 64 is to get a free ram upgrade, not really necessary
TommyC
No that' almost never secure, and highly improbable to do correctly.
pj
autofsckk: there is currently no stable 32 bit version of 7
and even if there were it would still not be possible to do what you want securely
Motoko
I highly recommend starting from a clean state.
Upgrading and changing core architecture is a bad idea. I'm experienced and I wouldn't even try it.
pj
the cost of overlapping two linodes if you take less than a week is only a few dollars, surely you can afford that?
TommyC
I did it once (albeit not on CentOS). It was a fscking mess.
I promised myself from that day forward I would just back up the files I need and just reinstall. Never again.
pj
just fire up a seonond linode, then start transferring your apps and data over one thing at a time. When it's done and everything is running on the new linode like you expect then you can shut down the old one.
cost will likely not be more than $5 extra than you're paying already.
TommyC
Also, ask yourself, do you *really* need the upgrade to CentOS 7 before support ends for it?
pj
why exactly do you feel you need to upgrade anyways?
autofsckk: ^^^^^
autofsckk
well i really dont know centos, i mean, i have been using linux for a long time, the thing is i tried to install odoo and i dont remmember right now what exactly i needed to install that i couldnt because of the version, i read for 3 days maybe trying to solve it but i couldnt
everything else is working just fine
pj
autofsckk: from what I can gather you need python 2.7 and postgresql 9+ for odoo, those are available for CentOS 6
python27 is available form IUS and newer postgresql versions are available from repositories maintained by postgresql directly.
autofsckk
yes, and i have both installed now, but python version is not the mmmmmm well i have 2.6 installed also but it is not the primary? i mean whe i tried to install odoo i got some errors from the python version but it is already installed
pj
well, the command for python27 is ... python27. There are likely several ways to make odoo run under python27.
« prev 1 2 3 4 5 6 next »