PDA

View Full Version : 3hr+ boot on AMD AthalonX2 with F7



cbetancourt
28th June 2007, 05:54 PM
Hello all,
I'm having a really strange problem...
On two machines (both are Athalon X2's on a foxconn n570sm2aa motherboard) I'm seeing exceptionally long boot up times longer that three hours.

I'll post exactly how long it took and other information on the machines actually come up. Any thoughts preliminarily?

Thanks

JN4OldSchool
28th June 2007, 05:59 PM
hmmm, I guess that rules out a hardware problem such as faulty PSU or RAM. I run an X2 without issue on an Asus board. Off the top of my head you might want to peek into your BIOS and see what state your cool and quiet is in. Try disabling (or enabling?) it. Other than that try doing a search on that board in here or in google.

leigh123linux
28th June 2007, 06:15 PM
Have you tried update the bios on the foxconn's ?

cbetancourt
28th June 2007, 06:38 PM
Not yet. I'll have to wait that till tommorow, when I can get my windows boot disk flash drive.

cbetancourt
28th June 2007, 06:39 PM
Oh, I forgot to mention, everything goes fine untill rpcbind starts, then the systems start dragging

leigh123linux
28th June 2007, 07:01 PM
Do you need rpcbind

if not

try

su -
system-config-sevices


and disable rpcbind for all runlevels ( uncheck all 3 boxes and save the change )

pitonyak
28th June 2007, 07:49 PM
The only time I have seen long delays during boot was when a process held up the boot so that it could hit an external site but it had a problem. For example:

I have two "system" partitions /dev/sda1 and /dev/sda2. I have FC6 on /dev/sda2 and it was my primary until they released Fedora-7. I then installed Fedora-7 on /dev/sda1. I forgot to change the label on /dev/sda1 from "/oldroot" to "/" (and something similar for /dev/sda2). The end result was that it tried (and failed) to do the right thing and a few things burped. Connecting to the time server failed, and it took a few minutes for the connection to time-out and fail so that it could continue. Imagine if it attempted to connect to multiple items and they all failed.

Can you place another computer on the line and sniff the packets to see what sort of activity is happening?

cbetancourt
28th June 2007, 10:44 PM
Ok, so total boot time came in around 5.5 hrs :eek: I can't disable rpcbind as one machine is serving nfs and the other is an nfs client.

As far as packet sniffing goes, what would you suggest to be on the watch for?

And the cool 'n quite settings didn't change anything.

Thanks for the suggestions thus far.

cbetancourt
29th June 2007, 02:18 PM
it was definitely rpbind. I tried an interactive boot just for grins, and when I didn't load rpcbind, I had no problems whatsoever. When I launch rpcbind by hand the system grinds back down to a halt

lmo
29th June 2007, 02:35 PM
That may indicate configuration problems with rpcbind et.al. I know nothing of this.

cbetancourt
29th June 2007, 05:58 PM
I'm starting to see problems with rpcbind on some of my other machines in the lab as well. I think I may start a thread in the server section. Would this be appropriate?

Thanks

mwette
29th June 2007, 09:06 PM
I may be having problems with rpcbind as well. When my server boots up under F7, nfs fails
to start. Once my machine boots, I can start manually ("service nfs start"). I'm wondering if
the order of bringing up network services is messed up somehow. Here is the error from my
messages file:


Jun 29 06:28:24 wetteserver kernel: Installing knfsd (copyright (C) 1996 okir@monad.swb.de).
Jun 29 06:28:24 wetteserver kernel: NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory
Jun 29 06:28:24 wetteserver kernel: NFSD: starting 90-second grace period
Jun 29 06:28:54 wetteserver kernel: portmap: server localhost not responding, timed out
Jun 29 06:28:54 wetteserver kernel: RPC: failed to contact portmap (errno -5).
Jun 29 06:29:24 wetteserver kernel: portmap: server localhost not responding, timed out
Jun 29 06:29:24 wetteserver kernel: RPC: failed to contact portmap (errno -5).
Jun 29 06:29:54 wetteserver kernel: portmap: server localhost not responding, timed out
Jun 29 06:29:54 wetteserver kernel: RPC: failed to contact portmap (errno -5).
Jun 29 06:29:54 wetteserver kernel: lockd_up: makesock failed, error=-5
Jun 29 06:30:24 wetteserver kernel: portmap: server localhost not responding, timed out
Jun 29 06:30:24 wetteserver kernel: RPC: failed to contact portmap (errno -5).
Jun 29 06:30:24 wetteserver kernel: nfsd: last server has exited
Jun 29 06:30:24 wetteserver kernel: nfsd: unexporting all filesystems
Jun 29 06:30:54 wetteserver nfsd[1870]: nfssvc: Input/output error
Jun 29 06:30:54 wetteserver kernel: portmap: server localhost not responding, timed out
Jun 29 06:30:54 wetteserver kernel: RPC: failed to contact portmap (errno -5).
Jun 29 06:30:54 wetteserver mountd[1889]: unable to register (mountd, 1, udp).

pitonyak
2nd July 2007, 07:07 PM
As far as packet sniffing goes, what would you suggest to be on the watch for?

I was thinking that you could watch the traffic from the machine that hangs. it is my guess that a request of some sort is sent and there is no reply. The tricky part is that it may never leave the machine so you may never see it.

If you know exactly what fails, it is easier to fix. I think, however, that you already know which service hangs; this is a very good start.

cbetancourt
3rd July 2007, 06:51 PM
I've figured it out!

In my nsswitch.conf file I had nis listed before files for passwd, shadow, and groups. When I changed the order, everything ran like it was supposed to.

Thanks to all for the input!

ronald.watson
7th July 2007, 04:10 PM
Someone mentioned upgrading the BIOS earlier in this thread which makes a lot of sense. Also confirming the BIOS settings for caching and shadowing are a good thing to check. Then I would try booting the system into the CLI (without the xserver) and see how long that takes noting all the daemons that are being loaded during boot. If it still takes a long time to boot (assuming there isn't an excessive number of daemons), then I would consider backing up any needed data, delete the linux partitions and reinstall.

Hope it helps.