logs archiveIRC Archive / Freenode / #centos / 2015 / August / 19 / 1
jeffspeff
one of our admins migrated a vm of centos to another vmware host. he only copied the vhd file and not the settings/config and when he booted it up he used the gui in gnome to configure networking. well, networking isn't working. previously the system used eth0 and now it's configured to use eht1 via ifcfg-Auto_Ethernet and Auto_eth1 config files. Is there a command or something to just undo all of the network manager configurations and just let me
edit the regular ifcfg-eth1 file? btw, i tried moving the *Auto* files and manually configuring via ifcfg-eth1 but it kept saying that the device eth1 wasn't found.
wolfy
jeffspeff: what centos release is it ?
jeffspeff
wolfy, if i remember correctly, it's 5 or 6
wolfy
make sure you delete the old MAC addresses from ifcfg-eth* and delete the file /etc/udev/rules.d/70-persistent-net.rules if it exists
then reboot
jeffspeff
wolfy, should i delete all the existing ifcfg- files? will they be recreated after reboot?
wolfy
nOPe
"make sure you delete the old MAC addresses from ifcfg-eth*" and "delete the file /etc/udev/rules.d/70-persistent-net.rules if it exists"
I did not say anything about deleting ifcfg-* files
jeffspeff
should i leave the ifcfg-Auto_* files?
wolfy
as you wish. the name of the file is not really relevant
jeffspeff
ok
there are currently 2 ifcfg-auto files and the old ifcfg-eth0 file.
         

wolfy
I for one would rm ifcfg-auto* and edit the old ifcfg-eth0
TrevorH
or rename the ethernet MAC address on the host for that guest to match the MAC in the correct ifcfg
wolfy
might lead to using the same MAC in 2 different VMs and make switches unhappy.
TrevorH
I took migrated to mean moved
jeffspeff
TrevorH, it was moved
wolfy
jeffspeff: that TrevorH's suggestion is also valid. make sure the relocated VM received from VMWare the same MAC address as before
TrevorH
which you can find inside the original ifcfg file
wolfy
s/that/then/
jeffspeff
ok, thanks guys
i'll take a look
Digipeng
centos 7, new pv ,and vg on lvm encrypted with luks setup and recorded in /etc/crypttab intended as a kvm pool. Do i need to add it to /etc/fstab?
the phsical volume that is
TrevorH
LVs have filesystems, PVs and VGs do not
Digipeng
Thought so
TrevorH
or in your case the LV has an encrypted mapping that would be the bit to mount
Digipeng
is it going to ask for a password on boot for the new pv?
TrevorH
something has to decrypt it
Digipeng
I hope so.
I guess ill reboot and take a look
TrevorH
I hope you have a way to boot if it goes wrong
Digipeng
It went fine so far
Yes! it worked
Bllacky
hello
anyone alive here?
Digipeng
!topic
         

agoodm
I am getting extremely infrequent list_del corruption...
http://pastebin.com/Ey6RSYZ8
Bllacky
!topic
agoodm
Last time I asked this everyone was screaming bad ram! so I tried to RMA the ram, however I was asked to perform 24 hours of memtest to verify the error, and wasnt able to fail the test... suggestions?
Bllacky, was that directed at me?
Digipeng
no that was for me.
agoodm
Digipeng, ah thats cool, lack of context there!
Digipeng
well @agoodm it would be nice to have more context to what your talking about. I know we would like to know centos version,and kernel version.
jeffspeff
TrevorH, so, that worked to get networking back up, however, i'm seeing 2 issues now. 1- editng resolve.conf and then restarting the networking service results in changes lost the top of resolve.conf says "#generated by networkmanager" even though i have "nm_controlled=yes" set in ifcfg-eth0.... the second issue is that if i issue the command "service network restart" it fails to start eth0 because it says the static ip is in use somewhere else, but
if stop the service then issue the start command, it works just fine.
agoodm
Digipeng, I cant remember how to double check the exact release I am running, however it was recently yum updated (27/7/15) kernel is errr
2.6.32-504.16.2.el6.x86_64
Bahhumbug
@alias add topic topic
centbot
Bahhumbug: Got it
Bahhumbug
!topic
centbot
Bahhumbug: Welcome to #centos | Don't Paste, use http://pastebin.centos.org | Only current releases (7.1.1503, 6.7, 5.11) are supported | Read http://wiki.centos.org/irc | Backports: http://tinyurl.com/r77l2 | Some thinking required | Be polite and considerate
agoodm
Digipeng, basically the story is that this machine has never been stable, however I have now had two crashes suggesting that the exact same bit flip occurred in the RAM, which seems really unlikely...
Digipeng
perfect. That makes sense.
!topic
lol
Bahhumbug
That didn't work for you?
Digipeng
@alias topic topic
@alias add topic topic
@whoami
Bahhumbug
Please stop now.
Digipeng
!whoami
Bahhumbug
@alias remove topic
centbot
Bahhumbug: Got it
Digipeng
fine centbot needs a run in with a buggszapper
well agoodm if it isn't the memory its the memory controller. not a good thing
bit flips in ram can occur esp not ecc ram
agoodm
Digipeng, I presume that its extremely unlikely to be a bug in the code that I am hitting then ?
Digipeng, its not ECC memory, however if you look at my backtraces its exactly the same bit flip as last time, it seems like too much of a coincidence
gaurdro
agoodm, have you checked if there is a bug filed about it?
agoodm
I have about 25 servers with a variety of differnt hardware configs, only two with ECC RAM and have ran into this issue 0 times before.
Digipeng
gaurdro
it's unlikely that you've hit a bug that someone else with your hardware hasn't already hit.
agoodm
gaurdro, Digipeng I did look on the bug tracker, but hadnt found anyone who had gotten the same bit flip twice...
newsoft
Hi all
So I have the following running trough netcat, for kvm transfer. Source :: dd if=/dev/vg/kvm165_img bs=8M Target :: dd of=/dev/vg/kvm165_img conv=notrunc,noerror bs=8M , this runs on 10gbps from both side, tested with iperf. When running the actual transfer, i can never reach speed higher than 50mb/s , and I can't identity the bottleneck.. any ideas ??
Digipeng
that would be your bs setting
newsoft
Source is soft raid-1 , that should easily reach >100mb read , target is hybrid-ssd raid-5 LSI , should write > than 50mb too
agoodm
newsoft, try larger/smaller BS setting
how do I double check which CentOS release I am currently running?
fenrus02
rpm -qf /etc/*-release |sort -u
agoodm
fenrus02, thanks
TrevorH
agoodm: actually the same bit flip sounds to me _more_ like dodgy ram than if it was not the same
agoodm
TrevorH, thats useful thanks, could you explain what the left hand side of the lines in the back traces stands for?
newsoft
agoodm ok , i will..
agoodm
newsoft, ensure you use the same bs each side
gaurdro
agoodm, when was the last time you updated the kernel and rebooted into it? I don't think that's the latest el6 kernel
agoodm
err should have been 27/7/15 - 2.6.32-504.16.2.el6.x86_64
TrevorH
agoodm: the "was" and "should be" values in both of those are not only identical but they differ by a single bit which makes it more likely that it's RAM IMO
agoodm
I've been yum updating and rebooting every time the machine has spontaneously rebooted. only recently have I had these two crash reports
to clarify though, the machine has not spontaneously rebooted recently
well not in the last 2-3 weeks... spontaneous reboots on average every 12 weeks or so
TrevorH
well looking at that I'd say RAM is your most likely culprit still but PSU could also be suspect
gaurdro
honestly, I'm with TrevorH that this sounds like bad ram. If this were my system and I could afford to have it out of service for a while I'd run memtest until I caught it in the act.
since that's what the vendor wants to rma it.
agoodm
TrevorH, the thing I thought that was strange was that the trace leading up to the issue was also identical... is my train of thought here somehow flawed?
TrevorH
yes, it's a specific check for corruption and the backtrace will be the same
« prev 1 2 3 4 5 6 7 8 9 next »