| View previous topic :: View next topic |
| Author |
Message |
imesg
Guru
Joined: 12 Jul 2002
Posts: 445
|
 Posted: Mon Jul 15, 2002 1:31 pm Post subject: random machine crash on network activity |
 |
|
I installed gentoo Friday and my machine has crashed more than 33 times.
I don't know if this is the proper forum to post this problem to, so if this is not it someone let me know.
My machine crashes or the routing table looses all entries. This happens randomly on network activity. I can click on a link on a webpage, or scp a file to another machine, or emerge xxx (crashes only during the download). It may be an hour or a minute between crashes, but it only happens during some kind of network activity. It may crash when evolution is retrieving mail via IMAP.
The ethernet is an onboard eepro100.
The motherboard is a Gigabyte P4 Titan series.
The kernel is gentoo-sources.
This machine worked reliably with another distribution as far as stability but I wasn't satisfied with the ability to make programs work so I tried gentoo, which I like except for the stability issue.
I know there is more information I can supply but I don't know what it is just now.
Thanks for any help,
Gene Imes |
|
| Back to top |
|
 |
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20067
|
| Posted: Mon Jul 15, 2002 2:26 pm Post subject: |
|
|
What are your CFLAG settings?
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
xming
Guru
Joined: 02 Jul 2002
Posts: 441
|
| Posted: Mon Jul 15, 2002 2:33 pm Post subject: irq |
|
|
check your irq setting:
cat /proc/interrupts
and try turn on APIC in your kernel config
xming |
|
| Back to top |
|
|
imesg
Guru
Joined: 12 Jul 2002
Posts: 445
|
| Posted: Mon Jul 15, 2002 5:09 pm Post subject: |
|
|
| kanuslupus wrote: | | What are your CFLAG settings? |
CFLAGS="-march=i686 -O3 -pipe"
This is from /etc/make.conf.
Thanks,
Gene Imes |
|
| Back to top |
|
|
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20067
|
| Posted: Mon Jul 15, 2002 5:13 pm Post subject: |
|
|
| imesg wrote: | CFLAGS="-march=i686 -O3 -pipe"
This is from /etc/make.conf.
Thanks,
Gene Imes |
Doesn't sound like that should be an issue.
| xming wrote: | check your irq setting:
cat /proc/interrupts
and try turn on APIC in your kernel config
xming |
Have you tried this yet?
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
imesg
Guru
Joined: 12 Jul 2002
Posts: 445
|
| Posted: Mon Jul 15, 2002 5:21 pm Post subject: Re: irq |
|
|
check your irq setting:
cat /proc/interrupts
root@go2:~>cat /proc/interrupts
CPU0
0: 160251 XT-PIC timer
1: 131 XT-PIC keyboard
2: 0 XT-PIC cascade
8: 2 XT-PIC rtc
9: 109 XT-PIC ohci1394, usb-uhci, usb-uhci
10: 415 XT-PIC eth0
11: 5843 XT-PIC aic7xxx, es1371
12: 5341 XT-PIC PS/2 Mouse
14: 5 XT-PIC ide0
15: 15 XT-PIC aic7xxx
What is the significance of this?
and try turn on APIC in your kernel config
[*] Local APIC support on uniprocessors
[*] IO-APIC support on uniprocessors
This is how I have had it.
Thanks for your help,
Gene Imes |
|
| Back to top |
|
|
pjp
Administrator
Joined: 16 Apr 2002
Posts: 20067
|
| Posted: Mon Jul 15, 2002 5:23 pm Post subject: Re: irq |
|
|
| imesg wrote: | | What is the significance of this? |
I have no idea. Hopefully xming (or someone else) will enlighten us 
_________________
Quis separabit? Quo animo? |
|
| Back to top |
|
|
xming
Guru
Joined: 02 Jul 2002
Posts: 441
|
| Posted: Mon Jul 15, 2002 7:38 pm Post subject: irq |
|
|
I was thinking you might have some kind of IRQ conflict, so I asked for cat /proc/interrupts to see which are in use by which device.
You network device is using 10 and not sharing it, that's good.
but do you really have 2 scsi controllers and only one ide controller?
APIC = advanced programmable interrupt controler, it's a new way of managing IRQ, your hw, bois and os have to support it to make use of it, modern hw usually do.
Now I noticed something else it's a eepro100 onboard on the Gigabyte? If it is then I had a lot trouble with it during heavy net load.
possible solutions:
- compile eepro100 as a modules instead of compiled in
- look for the eepro drv on http://www.scyld.com/network/eepro100.html because the one included in the kernel is very old
good luck
xming |
|
| Back to top |
|
|
xming
Guru
Joined: 02 Jul 2002
Posts: 441
|
| Posted: Mon Jul 15, 2002 7:38 pm Post subject: irq |
|
|
I was thinking you might have some kind of IRQ conflict, so I asked for cat /proc/interrupts to see which are in use by which device.
You network device is using 10 and not sharing it, that's good.
but do you really have 2 scsi controllers and only one ide controller?
APIC = advanced programmable interrupt controler, it's a new way of managing IRQ, your hw, bois and os have to support it to make use of it, modern hw usually do.
Now I noticed something else it's a eepro100 onboard on the Gigabyte? If it is then I had a lot trouble with it during heavy net load.
possible solutions:
- compile eepro100 as a modules instead of compiled in
- look for the eepro drv on http://www.scyld.com/network/eepro100.html because the one included in the kernel is very old
good luck
xming |
|
| Back to top |
|
|
Nitro
Bodhisattva
Joined: 08 Apr 2002
Posts: 661
Location: San Francisco
|
| Posted: Mon Jul 15, 2002 7:48 pm Post subject: Re: random machine crash on network activity |
|
|
| imesg wrote: | This machine worked reliably with another distribution as far as stability but I wasn't satisfied with the ability to make programs work so I tried gentoo, which I like except for the stability issue.
|
What was the other distribution? See if you can hunt down a .config from the other distro and use that as a starting point for compiling your kernel.
You don't get any kernel panic messages? Nothing in the logs, more specifically the kernel log?
_________________
- Kyle Manna
Please, please SEARCH before posting.
There are three kinds of people in the world: those who can count, and those who can't. |
|
| Back to top |
|
|
imesg
Guru
Joined: 12 Jul 2002
Posts: 445
|
| Posted: Tue Jul 16, 2002 1:57 am Post subject: Re: random machine crash on network activity |
|
|
| Nitro wrote: | | imesg wrote: | This machine worked reliably with another distribution as far as stability but I wasn't satisfied with the ability to make programs work so I tried gentoo, which I like except for the stability issue.
|
What was the other distribution? See if you can hunt down a .config from the other distro and use that as a starting point for compiling your kernel.
You don't get any kernel panic messages? Nothing in the logs, more specifically the kernel log? |
The other was Slackware-8. Too late to use the other .config. I installed gentoo over the jother. However I compile it with the same options except the ones recommended by the install document on this website.
No panic messages as I run gnome all the time and don't see the console. I can't ping and magic sysreq or anything but off/on. Good thing ext3, but after 33 times it does a filesystem check.
I think I like gentoo enough to try to find a fix.
There is no indication in the logs at all except a 'last' shows crashes.
Thanks,
Gene Imes |
|
| Back to top |
|
|
imesg
Guru
Joined: 12 Jul 2002
Posts: 445
|
| Posted: Sat Jul 20, 2002 2:15 pm Post subject: Re: random machine crash on network activity |
|
|
| imesg wrote: | | Nitro wrote: | | imesg wrote: | This machine worked reliably with another distribution as far as stability but I wasn't satisfied with the ability to make programs work so I tried gentoo, which I like except for the stability issue.
|
What was the other distribution? See if you can hunt down a .config from the other distro and use that as a starting point for compiling your kernel.
You don't get any kernel panic messages? Nothing in the logs, more specifically the kernel log? |
The other was Slackware-8. Too late to use the other .config. I installed gentoo over the jother. However I compile it with the same options except the ones recommended by the install document on this website.
No panic messages as I run gnome all the time and don't see the console. I can't ping and magic sysreq or anything but off/on. Good thing ext3, but after 33 times it does a filesystem check.
I think I like gentoo enough to try to find a fix.
There is no indication in the logs at all except a 'last' shows crashes.
Thanks,
Gene Imes |
For anyone interested, the fix was to get a 3Com 3c905b pci ethernet card and use that instead of the on motherboard Intel Ethernet Express Pro 100.
Caio
Gene |
|
| Back to top |
|
|
bludger
Guru
Joined: 09 Apr 2003
Posts: 389
|
| Posted: Tue Jul 29, 2003 4:44 am Post subject: |
|
|
| I have also been having this problem lately. I have a Realtek 8139 lan card. I thought that it was my nvidia graphics card playing up again, but I have since slowed down the clock and it crashed again when doing an scp. I am currently running with gentoo-sources_2.4.20-r2. I suppose I will upgrade to r5 and see if it happens again, but I am not hopeful. |
|
| Back to top |
|
|
Janne Pikkarainen
Veteran
Joined: 29 Jul 2003
Posts: 1143
Location: Helsinki, Finland
|
| Posted: Tue Jul 29, 2003 8:12 pm Post subject: |
|
|
For all the Intel Ethernet Express 100 (Pro) users: if you encounter sudden & random lockups during high network activity, please make sure you are using e100 driver, not eepro100. So do a lsmod and see what driver is loaded or check out your kernel config.
That did the trick in my case and is documented at IBM site, anyway. |
|
| Back to top |
|
|
|
Other Things Gentoo
