-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Scott Ehrlich Sent: Sunday, August 05, 2007 12:32 PM To: CentOS mailing list Subject: RE: [CentOS] CentOS vs RedHat packages
On Sun, 5 Aug 2007, Ross S. W. Walker wrote:
-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Scott Ehrlich Sent: Saturday, August 04, 2007 8:43 PM To: CentOS mailing list Subject: Re: [CentOS] CentOS vs RedHat packages
On Sat, 4 Aug 2007, Shawn Everett wrote:
<snip>
The basic issue I have encountered is, after freshly
installing RHEL5
64-bit Server on a Dell Poweredge 2950, everything is fine. After logging in for the first time, I configure services as I want via system-config-services. But after doing so, and rebooting, at first the filesystem claims there is a system error and prompts for the root password for a fsck, or control-D to reboot.
Scott, do you remember what the system error was at boot?
It may be that config-services turned off an essential
system startup
service.
I found 'chkconfig' much more reliable and easy to use on
the CLI than
config-services in the GUI.
I'll check that out - thanks!
Also what method did you use to reboot your system after
config-services?
I forget - either reboot at the command window, or reboot from the gui. Never a "hard" reboot.
Did it hang while doing the shutdown that forced you to hit the power button?
If it doesn't have Internet access, does it have any network access, cause if DNS isn't working properly and /etc/hosts isn't fully populated there can be hangs while it timeouts doing name-lookups and I have had some admins think that the machine is frozen and do a hard power-off and power-on then wonder why their file-systems were corrupted.
I opt to reboot, and the system then complains of python errors, but there are no more system errors. But, when I revisit system-config-services, I immediately experience python errors.
Possible file-system corruption in and around where the
python libraries
are since they were last used... Maybe due to a dirty shutdown?
I always go for a "software" shutdown where/when possible - never power cord nor power switch.
To try and potentially help resolve this problem, I recently updated the BIOS and some of the firmware. I then performed a fresh install again. I then downloaded RH's Enterprise 5 Supplement ISO, along
with Dell's
OpenManage ISOs for Enterprise 5. I'm hoping something among this back of changes will help solve the python configuration problem.
Did you install the latest dell drivers for your controller
(Perc 5/i?),
the drivers for these controllers are updated/fixed frequently.
I let RH install its own drivers. I will probably then simply reinstall the OS again, this time with linux dd to utilize the Dell PERC drivers.
Linux like any OS has basic drivers to get it up and running, but those drivers tend to be a little out of date and they could have some critical issues that appear when run under load.
Best practice is always to install the latest hardware drivers from the manufacturer after/during install, especially for storage and network (video for workstations).
Unfortunately, I was unable to get the PERC firmware to successfully update with Dell's utility. I may ask them for a resolution on that, unless I can overcome it another way.
Make sure compat-stdc++-33 is installed as the Dell installers use the older c++ libs.
Making sure the firmware is current is also critical because some of Dell's initial firmwares have some serious integrity flaws.. :-(
If not, I may opt to see how well CentOS 5 handles things.
It will unfortunately handle them exactly the way RHEL
does. As it is
RHEL...
I'm mostly interested in the PostgreSql and Java pieces.
And a stable and reliable system to run them on I hope. ;-)
Of course.
If it comes down to it, I could always place a trouble call with Dell and RedHat and see if either of them has a helpful solution.
And the forecast calls for pain...
Anyone on the list experience the same kind of problem, and find an answer? If so, what was it?
At this point it really can be anything, I would start by
getting the
latest PERC driver from Dell and installing it, uses dkms to keep it current on updated kernels. You will need cc, binutils, kernel-devel installed before hand for it to fully work right on the initial install.
Always pre-installed - never know when they will be needed...
Good, cause a lot of people install the dkms and then the dkms-driver and think it is installed, but it wasn't able to compile and install the driver because the build system wasn't in place during the install.
A 'dkms status' will tell you the status for each dkms managed driver for each kernel installed on the system.
dkms is especially handy when you have a bunch of Xen domU kernels installed too as it can manage drivers for those too.
The server is not connected to the Internet, and likely never will be - strictly out-of-box CDs.
Should not make a difference, except you will have a hard
time getting
the latest bug-fixed and security-fixed versions of the system.
I would look into finding a way to get this puppy to get the latest updates, either via proxy server or internal update server.
I've got other methods.
I'll try your suggestions when I return to work tomorrow.
Thanks!
Scott
Good luck,
-Ross
______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.