To avoid this error, upgrade glibc first in a separate run:
Nice! So all the systems that autoupdate won't have the logic :)
Glad I don't do that...
jlc
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
Whatever I do - it needs to be convincing enough to make the PCI
compliance guy tick the box.
Eset has a current linux client, though their product *AND* support
suck the biggest one.
https://www.icsalabs.com/icsa/product.php?tid=dfgdf$gdhkkjk-
for more
HTH,
jlc
I am trying to follow a solution that states edit apache's libtool and
change the following option to yes: build_libtool_libs=no which begs
the question, how? :) Anyone care to show me the light?
Thanks!
jlc
___
CentOS mailing list
CentOS@centos.org
I am trying to follow a solution that states edit apache's libtool and
change the following option to yes: build_libtool_libs=no which begs
the question, how? :) Anyone care to show me the light?
Found it, /usr/lib/apr-1/build/libtool, but it was set to yes:/
If you'd tell us what it is that you're doing we can help out, or
explain any one of a hundred ways that you're doing it wrong. Either
one, really.
I was tearing my hair out trying to get mod_auth_ntlm_winbind working with
apache so I could later use adLDAP/Dokuwiki to migrate all our docs out
mod_authz_ldap definitely can use groups or users to authenticate though
to be honest, I am only authenticating to OpenLDAP and I do see some
references to authenticating to active directory in the documentation.
Yeah, I have LDAP working w/ groups using Dokuwiki's built in LDAP auth
backend. It
Can I use 'linux rescue' to fix that OS up to boot it? The kernel
panics in its current state (because the hardware architecture is so
strikingly different). What is the methodology of fixing the kernel
in this circumstance?
You likely don't have block device modules for whatever controller
I am trying to setup moinmoin and my apache virtual host config looks like this:
VirtualHost *:80
ServerName www.example.net
DocumentRoot/path to doc root
AddHandler cgi-script .cgi
Alias /moin_static181/ /usr/local/share/moin/htdocs
ScriptAlias / /path to
How can I do this w/o restarting services or rebooting?
Thanks!
jlc
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
If I understood well what you are looking for, to manually rotate your
logs, you could launch logrotate as follows:
logrotate -f /etc/logrotate.d/service
That was the ticket!
thanks!
jlc
___
CentOS mailing list
CentOS@centos.org
When I run this, I have ssh,http checked and have added Kerberos and squid
manually
below, but when I save it I still get a port 88:tcp added to the
/etc/sysconfig/iptables
file?
What can I use to track what is forcing it to get added? The box is a current
CentOS 5x86 server. I was running
I was running apache on it with it set to 88, but it has
since been moved back to 80.
Since the box is a member of an ADS site, the Kerberos threw me, it
took the 88 we entered, and I forgot the tui converts your numerical
entry to the text equivalent.
Sorry for the noise...
jlc
I need to review a logfile with Sed and cut out all the lines that start with a
certain word, problem
is this word begins after some amount of whitespace and unless I search for
whitespace at the
beginning followed by word I may encounter word somewhere legitimately
hence why
I don't just
So how can I find out what Centos knows about the system temp and fan state?
LM_Sensors does this...
let me know what you do to get it working on your rig :)
jlc
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
The regex you want is ^[[:space:]]*word
Wow, thanks everyone for the help! How does one modify this to also knock out
lines that *must* have whitespace followed by a number [0-9]? I can do it using
^[[:space:]]*[0-9] but it also takes out lines w/o whitespace that begin with
numbers?
I have to
to match one or more, use + instead of *.
* matches 0 or more, + matches 1 or more.
Thanks!
I have to buy a book on RegEx's and Sed :)
http://www.gnu.org/manual/gawk/gawk.pdf
(G)awk is pretty sh!t hot where I work; however we've extended it a
bit. :)
So gawk does all that sed does and more?
I read something in a forum, that said the CPU temperature can go to
100 C., in a few seconds, but I now wonder if that is true or not.
Yeah, turn on a machine w/o a heatsink :) In a couple of seconds it will
get so hot that you can't touch it and the internal safety threshold on the
proc will
Thank you Joseph, that did the trick! After studying the man page for
mkinitrd, I also used --preload ahci to make sure the ahci driver was
the first one to look for drives.
Good to know,
I think if you had the modprobe edited via a rescue disc and you chrooted
in to the dead system the mkinitrd
So I'm wondering what config files I need to tweak, to help the
already installed system use the new ahci controller? Previously,
it appears to have been using the piix module to access the drive with
the root fs on it.
So, you need modprobe.conf to reflect this as well as a new initrd.
On a
You use kinit before joining the AD to test AD auth. That is, you want
to be sure your linux side is configured properly to get a kerberos
ticket in the first place. If you're able to get one, you should be to
join the domain.
Ranbir,
Yeah it's been working ever since but there are some errors
When you use kinit to acquire a Kerberos ticket prior to
joining a Win2k3 domain with net ads join -U upn is that
ticket merely used for the join that follows? When it expires
does this mean anything?
Thanks,
jlc
___
CentOS mailing list
Under CentOS 5.2, is it possible to mount an iscsi filesystem/partition as
rw for root, but ro for users?
Possibly, but iSCSI is not a file sharing protocol. You can only mount cluster
FS's with multiple ini's or when root changes the FS while another client has
it mounted funny stuff happens
Any advice?
Only that PCIe in my experience needs a quality riser for signal interference.
If this is a server running anything important, I would be leery about
engineering
my own setup from generic components.
YMMV,
jlc
___
CentOS mailing list
I assume that based on your last message to the list of is there
another way to handle redirects other than installing an Apache
server and now this, there is no other way?
I only ask because I am now setting up squid in my spare time and this
is the next task to handle?
Yea, it doesn't look
I have a squid proxy running transparently, so in my firewall script
I run the following fairly early:
iptables -A PREROUTING -t nat -i $LAN -p tcp -m multiport --dports 80,443 -j
REDIRECT --to-port 3128
This is a multihomed server so after this change the masquerading was
removed (as only web
Makes sense to me.
Yea, I just don't know technically speaking where the -m mac should appear, in
the POSTROUTING line, or the first FORWARD line. Ultimately I would only
masq'ing to be done for this one device on port 443.
Is the host that you are wanting to bypass your proxy on the same
I don't have a web server on a machine running squid and
squidgaurd and don't really want to add one. Is there a way
to do something other than redirect in an acl, such that users
could see why the page they are trying to get to is not available?
Thanks!
jlc
I have this cgi file running on CentOS 5 with Apache 2.2.3 and although it
executes most fields are empty, a typical block by squidGuard yields this:
Supplementary info :
Client address = 192.168.0.5+clientname
Client name =
User ident =
Client group
I can't find a Linux driver for it except one that appears to require me
to recompile a custom kernel. I find it on the Marvell website:
http://www.marvell.com/drivers/search.do
No, you make a module with that downlaod. I also have the unfortunate luck
to have a few systems with various Marvell
I've also been using Untangle (untangle.com) and just love it.
This machine is nearly stock with all the nat/firewall done in a simple
hand written script, it also serves as an Asterisk PBX so I couldnt use
an appliance.
I'm not sure if the latest has all the features OP is seeking, but I've
yumlist or yumex are not as good as rpmfind, as i don't
run yumex/GUI on any server.
How does #yum list name not work *if* the repo is added to your list?
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
How can i get of that LVM and clone it to the new SATA drive?
Well, you need modules in your new initrd for the sata controller.
It looks like you have everything else in order.
FWIW, I do softraid under LVM no problem and I like the flexibility
of LVM so I always use it now.
jlc
Is that how rpmfind works?
Don't know _exactly_ how I searches, but I think that point is mute.
http://rpmfind.net/linux/RPM/ What packages are available for CentOS and third
party repos
without having the .repo files installed?
Sorry, I don't understand. Add the few repo's you know you'll
That could be the drive order confusion I mentioned.
Good point,
What happens if you yank the IDE drive and attempt to boot?
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
I try'd in different order. Disconnect the IDE drive, change only the
bootorder etc. Nothing, everytime the same message. Grub can boot it,
its running into kernel and one line before init is running kernel
panics. Seems like the kernel could not find the root device, the swap
is afaik not
Mobo is: Intel D945GCLF Little Falls
Iv'e try'd mkinitrd with mounted /sys, /proc and /dev from the live cd
without any special parameters, so my think was, sata is visible and
should be enabled by default. The LVM is not longer included in the new
initrd because the lvm messages are gone on
There seems to be a lot of fanboy affinity around ESXi - and with the
fact that its 'available' off the shelf, zero cost up front. however to
make it do anything you still need to buy into vmware tools.
Huh, Tools are free? I do _a lot_ of nothing apparently without what
I assume you meant,
I've been repeatedly told ( including by people @vmware ) that you need
the VI-client in order to get a management interface on ESXi, which
neither runs on Linux nor is freely available.
Am I being lied to ?
yes, VI is free. It does not run on Linux though which sucks, but same for
XenServer's
offers medium to low performance compared to other
similar products
I get it, you *hate* Microsoft and Windows and ... That's cool. But before you
make claims about facts (not opinion, which is very valid as I respect your
personal choices to be good for you) you should verify those. Vmware is
I'd be happy to have you over my place and we can do some real world
performance testing in server roles
I would actually love that (just for the sake of learning) but me thinks were on
opposite sides of the pond:)
btw, as might not be clear to some people, I dont do Windows
hosts/guests so
Why not give kvm a try
i am getting a BSOD on shutdown, but so far it is not bothering
anything afaic tell.
You're not exactly making a good pitch bro :)
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
I am trying to clone harddrive
eg: centos, freebsd, openbsd.
any ima ge software suggestion for different os
Clonezilla is pretty versatile but has limits going from disc's with
different sizes. I will use the app best suited to the fs in question.
jlc
I have an issue with only a few domains sending mail to a company's exchange
server.
Topology is as follows:
Internet - PIX - logical interface / vlan - HP switch (interface tagged into
applicable vlan) - ESXi server - vm (ASSP/CentOS Postfix MTA) - HP switch
(out of vlan) - physical HP server
How does one redirect output of a command such as 'make' to a file
_and_ the console?
Thanks!
jlc
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
This is either a problem in the RPMFORGE repo ... OR ... a problem with your
yum-priorities settings blocking an rpm from RPMFORGE that you need.
Yeah, looks like rpmforge has an issue atm:
# yum --disableplugin=priorities whatprovides */libFLAC.so.8
Loading fastestmirror plugin
Loading mirror
In the future, when upgrading the kernel, will future initrd's be built with
my current modules or will I have to manually create new initrd's after each
update?
Good question that I would love to know as well. I thought
/etc/sysconfig/mkinitrd handled
this, but not ripping apart the srpm of
I have a project that I need some hardware pointers for. I need to build
some Centos appliances (dedicated boxes to do one thing only). Target
cost is under $250/box.
Given the rest of the requirements, I would say something like:
http://www.mini-box.com/M200-LCD-Enclosure
Find a distro more
I have some pm files I want to add, these seem to be custom modules delivered
with a package I am trying to install. I don't want to run its script as it
manually installs many modules available in base so I will simply yum those
in. How can I add the pm files, I assume I have to update something
Why do you try to filter outbound connections at all? If something makes
it on your machine the first thing they will do is drop your rules.
You imply the *only* reason for outbound filtering is stop a hacker. In some
environments it serves as an additional layer of protection against other
What am I missing? Is it odd to have two SATA controllers for a
single disk drive, or is that at all relevant?
You don't, it would be odd :)
Actually, we do - it is configured to have two SATA controllers in it.
Heh, Actually you don't have two SATA controllers for a single disk drive
but you
That brings me to ask those of you that use command-line burning if you know
of any way I can close this disk. I suspect the files on it are ones that I
don't have elsewhere.
Use cdrecord to fixate the disc (--fix).
jlc
___
CentOS mailing list
I'm not familiar with doing it this way, so could you be so kind as to give me
a complete command? I guess that I could work it out from the man page, but it
would take a lot of trial and error, and maybe getting something wrong would
be risky.
Off the top of my head:
$cdrecord -scanbus
See
Should we go ahead and reset to AHCI and reinstall (again)? Or will
enhanced be enough for the long run?
Enhanced/AHCI for SATA drives under Linux.
Good info here:
/var/log/messages.1-Nov 1 18:19:00 Centos5 kernel: hda: status
timeout: status=0xd0 { Busy }
/var/log/messages.1:Nov 1 18:19:00 Centos5 kernel: ide: failed opcode
was: unknown
/var/log/messages.1-Nov 1 18:19:00 Centos5 kernel: hda: no DRQ after
issuing MULTWRITE_EXT
/var/log/messages.1-Nov 1
Listing the RPM packages you need on the RPMforge mailinglist would be
sufficient to have them added (where possible).
Dag,
I can't tell you how much I would appreciate that though I don't envy
this task. Aside from ASSP 2.0 needing many, it requires newer versions
of modules that come in Perl
cpan2rpm is getting relatively old, compared to some of the newer
things that can be done with the CPAN and RPM build procedures. The
latest version isn't available on its sourceforge site and the
arix.net FTP link requires a login
I yanked it off arix.net?
All this effort is to build a CentOS
Just pvcreate the whole disk and forgo partitioning it. Then create a
vg out of it and start creating lvs.
Hey Ross,
I thought it was best practice to create an LVM partition such that the
disk could be recognizable under all circumstances such as if the volume
was moved? Is that not really
I'm not really sure what the difference would
be having the physical extent size as 64 MB versions 4 MB.
It's the smallest allocatable size of any PV you can give to an LV.
Think of it like Allocation Unit Size. Also as pointed out above some
limitations arise in different versions of LVM. RH
I vaguely recall reading that fdisk wasn't suitable for partitioning and
wonder if I shouldn't be using partd instead. I am also wondering if I
should use lvm or just mkfs to create the filesystem. Anyone have
suggestions before I blunder in?
fdisk can't do GPT which is what you need for
XP Home was purposefully crippled by MS so it lacks networking features
in XP Pro and Win2k Pro.
Yawn...
Are you informed well enough to know how it was crippled as you say?
I think not, it may not be a member of a Windows domain. That's it.
It costs less. That's fair?
I wonder if you apply your
I have never setup a Samba server, and the recent thread here prompted to
get on this I need to migrate two XP machines from using iSCSI to Samba.
I thought the + preceding a username in the smb.conf file for a
share definition told Samba to use the unix username? Unless I actually
use swat to
Why would you put user names in the smb.conf file?
The server is set to user hence the need to suggest what user's can
access the share, right?
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
I've always had to use
smbpasswd -a unixusername
then give them a password for smb use... the catch-22 is that SMB uses
a completely different password hash algorithm than Unix/Linux
/etc/passwd(shadow) passwords, so you can't use the one for the other.
Its much much easier if you use a
you know individual Windows DOMAIN\usernames can be added to the local
Administrators group on a workstation?
yeah, but like I said, I didn't want a logged in user (can't even enable a
locked
screen saver) to be accessible by someone. I could create perms denying this
user
explicitly etc but
I used to do manual installs so I would set the timezone each time.
I have now migrated to kickstart and I wish to use a command line (at
boot) parameter
to set the needed timezone.
What's wrong with the KS Timezone command? Why do it via cmdline or w/ symbolic
links??
I always used IET on CentOS but I have a quick server to setup to backup some
VM's, then migrate to DAS
on another server and the easiest way for me will be to accomplish this via
iSCSI vmfs stores. I figured I would
just use the ini that ships with CentOS but have never used it against ESX,
I have CentOS v5.2 running as an iSCSI target server and multiple ESXi servers
as the initiator with VMFS. This works great! The tgt daemon that comes with
CentOS is simple to support as long as you know how to use tgtadm to setup
the targets. I modified the tgtd init script to allow me to use
The package xorg-x11-libs doesnt exist
# yum list | grep -i xorg-x11-libs
No, but some simple digging into what it *could* be
might lead you to this:
[EMAIL PROTECTED] ~]# yum list | grep -i x11 | grep -i lib
libX11.x86_641.0.3-9.el5installed
libX11.i386
I have to believe that you have heard of Wikipedia:
http://en.wikipedia.org/wiki/Crusoe_processor
I just did not know what to use for my googling to get to this. Thanks.
This wiki page even lists the OQOs as examples of crusoe processors.
Now in terms of actually running Centos on this
On the quotes we've been getting, XenSource would cost half of what
VMWare ESX would cost (not the stripped down free download obviously).
We are a cash-strapped organisation, so the difference is significant.
Download the free ESXi and give it a whirl. Thats so much good software for
free!
jlc
hey -- I have and LIKE that book ;)
I'll second that. I have a few and this one is hands down, the best to date.
jlc
___
CentOS-virt mailing list
CentOS-virt@centos.org
http://lists.centos.org/mailman/listinfo/centos-virt
My boss seems most interested in a paid XenSource enterprise solution,
which apparently runs on CentOS?
If you are willing to pay, xensource has nothing on vmware. I have ran both
and vmware is far more polished and enterprisable IMHO.
You can't even do a snapshot in xensource yet. Its far to
umm ...
xm save (domain-id) (state-file)
... misc copy actions
xm restore (state-file)
You are asserting the Xensource lacks what the CentOS supplied
xen has? wow
How many of those can you do? Can you nest them?
Enumerate the steps to add an LV backed ISO sr.
Oh, reactivate the LV
I've got three Promise SATA300 TX4 controllers running in Centos 5
boxes without any problems.
Thats proof 1 in a million isn't impossible :)
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
Unable to look at what's on dbus
A syslog message that coincidentally popped up while
you were doing a yum install?
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
Maybe so, I've never seen yum do that.
Again, yum didn't. It was a message that simply interrupted the
stream of text that was being output to your console. Likely syslog
did this.
jlc
___
CentOS mailing list
CentOS@centos.org
Not a laptop it is a server with dual nics mtu set at 1500, 1 nic directly to
the cable modem the other to my lan.
MTU has nothing to do with laptops, it has to do with TCP/IP.
Aside from a previous posters duplex setting (good catch on his part) its
possible that MTU is an issue. Some ISP's do
I have a server with dual nic's running centos 5.2 1 nic connects to the
cable modem, which has 15m d and 2m up, but speed tests barely come back at
5mb down including road runner's internal speed test. I hook up my laptop to
the same modem, same port, using the same cable and I get 17mb down. I
The Dlink DNS-323 looks exactly what you are asking of
What a procedure to hack that thing!
The problem I see with going the all-in-one NAS route is that down the
road, there's always some function you'd like to add - but you can't.
You've hit the limitations of the box.
That's why I want to
I've /never/ seen RHEL/CentOS or any of its predecessors renumber
ethernet ports on a working system..
Yeah, I have never seen it renumber either?
I've seen it number them
backwards, such that eth0 was the port labeled '1' outside the chassis
and eth1 was port '0', but it was extremely
I get a message that pops up Could not allocate requested partitions
press OK to reboot.
Got the clearpart in there?
jlc
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
This is a VERY common problem. I see it all the time.
I also can confirm this *very* typical behavior.
Paolo,
Remind me what the issue is that prevents you from working through
this based on the known existence? Its been muddied somewhere through
the thread...
jlc
I am trying to setup a series of vmware esx 3i systems to be provisioned by pxe.
Since you can pass the extracted config to the hypervisor in the pxe append line
as a file on the tftp server, I presume I can use my CentOS DHCP/PXE server to
hand
out assigned ip's to specific hosts. Thus, I need a
I haven't tried esx 3i myself yet, but kickstarting ESX 3
was similar to red hat kickstart with anaconda and stuff,
I did all my configuration by hand afterwards because of
the trunked/tagged network ports which aren't compatible
with PXE, and I surely didn't want to risk an automated
install to
Network autoconfiguration failed, most likely there is not
a compatible driver for the network card in your system.
Paolo,
You are using multihomed systems, I remember a problem with pxe booting
whereby the install switched to the wrong nic and setup may fail as a result.
How are you statring
Is there a way to nice the IO on a process such as dd?
If not, what could be a way to control the IO level of
such a process from bogging down a server to severely.
Thanks,
jlc
___
CentOS mailing list
CentOS@centos.org
WHat do you mean?
Do you boot from one of the NICS, or a CD, or what?
If you boot from a NIC via a TFTP server, you may be encountering the issue
I describe which simply needs a kernel arg to fix you up.
jlc
___
CentOS mailing list
CentOS@centos.org
I boot via NIC0 and my configurations are as follows:
Yeah, that's what I thought.
Have a look at:
http://www.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5.2/html/Installation_Guide/ch-pxe.html
http://kbase.redhat.com/faq/FAQ_80_531.shtm
The latter might shed some light. I am thinking you
nice doesn't really do anything with respect to I/O.
Yes I tried it and it never made a diff from one end of
the spectrum to the other:)
The best way to control I/O in this manor is to physically
isolate it from the rest of the system(be it on a different
controller connected to different disks
I need to replace a slow HP MSA20 that's a bottleneck right now (We are an HP
shop).
HP's only option with 3.5 SATA drives is stuck at SATA I behind a SAS
controller.
Dell makes an MD1000 that claims SATA II speeds behind a SAS Perc 5/E (Don't
know anything about dell).
Anyone got any real
If that doesn't do it for you then maybe choosing a different
scheduler then cfq can help. Something like 'deadline' may
work better for the workload.
AFAIK ionice will only work with the cfq scheduler for now.
Appreciate that info, I have just been reading about the difference but
cant say I
I assume you're talking about the MSA60?
Yup
We have several of those installed
and haven't seen any bottle necks there. Even the fastest SATA drives fall
short of the 150MB/sec max SATA I throughput. And if you could push the
1.5Gb/sec on the SATA side, then the SAS side with a 4x port
Sorry for hi-jacking this thread, but it's very interesting. What's the
difference, apart from the speed price between SAS SATAII?
SAS is more enterprise geared, relating to speed and MTBU (probably a load of
crap on that point)
and SATA is cheaper and does not perform as well, supposedly
There was a thread about the MTBF of various types of drives, linking
to articles about the experiences of Google.com, here in this mailing
list, probably 1 or 2 years ago. As I recall, they found the MTBF
between different types of drives to be about the same?
That's exactly what I was
We should be talking live. Why don't your join the #centos-social on freenode
so we can chat real time?
Robert,
Just got back from my trip and reading that Tutorial, it went on to state
what I now find to be two distinct opposite thoughts. Its says at
why would md1 not be in sync after an install.
Jerry
I just did an Asterisk setup with a similar MDRaid setup and noticed the same
thing after a few quick restarts, I should have checked the status of the
(re)build/sync
but as Romeo suggested it makes sense.
jlc
I'm wondering if anyone out there is doing tap:aio disk devices on a Xen HVM
vm with any success.
I am on Xen 3.3.0 and have not made this work for HVM iether, I have been
following those threads to...
jlc
___
CentOS mailing list
CentOS@centos.org
Postrouting is the correct one. After everything is routed it is MASQ before
leaving the interface.
That makes sense, but I am stuck at making the transition, should I simply
pull the port specifications from my third line in the FORWARD chain to the
first line in the POSTROUTING chain?
No port or packet types are needed as everything needs to be MASQed.
Sorry, I wasn't clear. I only want ports 80/443 translated for internal clients
so I do need a rule of some sorts.
Thanks!
jlc
___
CentOS mailing list
CentOS@centos.org
You can do this in /etc/modprobe.conf (or /etc/modprobe.d) OR you can
create an /etc/rc.modules (for legacy support) OR you can add your
bits to /etc/sysconfig/modules/foo.modules.
Jim,
I did assume modprobe.conf was the correct place and tried a line which didn't
work. man modprobe.conf shows an
601 - 700 of 972 matches
Mail list logo
