Hello Everyone,
Ever since installing my FreeBSD machine I have been plagued with an issue. The installation went beautifully, during which I used bsdinstall to configure my re0 (eth0) interface. Every piece of software I have installed I compiled from source. I also figured out how to get ntpd running and automatically started at boot time via rc.conf and afterwards installed Xorg, followed by Gnome2 which took a good part of 17 hours to build.
I first noticed the issue the first time I attempted to boot up my machine without an internet connection. Everything went fine except at the end of the boot dmesg info, you could see that ntpd was freaking out as it couldn't connect to its NTP server, and it kept trying. I ignored it and started X. After about 3 minutes the system started slowing down, as if my computer was literally dying. It slowed so far that I would move the mouse on my laptop and the mouse would move 15 seconds later and got progressively worse until the system was completely unresponsive.
I rebooted expecting it to happen again, and checked top and realized it was not a particular process causing this, if I tried moving the mouse moused would shoot up to 80-99% of the total CPU usage. If I stopped, Xorg filled its place. Basically everything the computer tries to do shoots up to 80-100% of the computers total capacity, until finally the computer completely locked up.
I first thought the problem may be caused by ntpd, so after experimenting with killing it right after logging in to the machine (which did not work), I just disabled it altogether in favor of NTPDate. This changed nothing. I then thought to myself maybe perhaps it was cron in conjunction with ntpdate causing the issue. Killing the cron and NTP processes as soon as I log in still does nothing. Needless to say if I boot the computer up without an internet connection the computer may last a couple, even 4-5 minutes but every single time I can see it coming when moused or gnome-terminal starts using 10%-20% of my CPU. It rapidly gets worse until the computer is unusable. I cannot switch to a virtual console to stop X, nor shut the machine down. In fact I am lucky if I get it shutdown when I first notice it happening as before the shutdown procedure can complete its already usually dead.
Seeing as it is DIRECTLY related to having an active internet connection or not, I started thinking of everything the system loads automatically which requires an internet connection to function. Unfortunately I never get past ntpd/ntpdate. What's really baffling is that kill -9 'ing all of those process, in fact one time I did ps aux -c | grep less from the virtual console before even starting x, and killed almost everything not critical to the machine. Things like adjkerntz or cron. I figured the atrun or adjkerntz commands cron attempts may be causing it, but killing them should return the memory to the CPU, however this is literally as though the memory leaks out of it and cannot be restored.
Does anyone have ANY ideas as to what may be causing this, or even an educated guess at a list of possibilities. I have a feeling that it's a process creating a memory leak recursively, for a corefreebsd FreeBSD process to be causing this seems to me would be out of the question however so I am at a complete loss, I can provide whatever info on my system necessary.
My System:
Laptop with about 1.5G Memory and 1.8G processor AMD Chipset
FreeBSD 9.0 Release
Xorg - > Gnome with no GDM management, however the exact thing happened before i built gnome and was still using twm with no XDM management.
/etc/rc.conf:
Ever since installing my FreeBSD machine I have been plagued with an issue. The installation went beautifully, during which I used bsdinstall to configure my re0 (eth0) interface. Every piece of software I have installed I compiled from source. I also figured out how to get ntpd running and automatically started at boot time via rc.conf and afterwards installed Xorg, followed by Gnome2 which took a good part of 17 hours to build.
I first noticed the issue the first time I attempted to boot up my machine without an internet connection. Everything went fine except at the end of the boot dmesg info, you could see that ntpd was freaking out as it couldn't connect to its NTP server, and it kept trying. I ignored it and started X. After about 3 minutes the system started slowing down, as if my computer was literally dying. It slowed so far that I would move the mouse on my laptop and the mouse would move 15 seconds later and got progressively worse until the system was completely unresponsive.
I rebooted expecting it to happen again, and checked top and realized it was not a particular process causing this, if I tried moving the mouse moused would shoot up to 80-99% of the total CPU usage. If I stopped, Xorg filled its place. Basically everything the computer tries to do shoots up to 80-100% of the computers total capacity, until finally the computer completely locked up.
I first thought the problem may be caused by ntpd, so after experimenting with killing it right after logging in to the machine (which did not work), I just disabled it altogether in favor of NTPDate. This changed nothing. I then thought to myself maybe perhaps it was cron in conjunction with ntpdate causing the issue. Killing the cron and NTP processes as soon as I log in still does nothing. Needless to say if I boot the computer up without an internet connection the computer may last a couple, even 4-5 minutes but every single time I can see it coming when moused or gnome-terminal starts using 10%-20% of my CPU. It rapidly gets worse until the computer is unusable. I cannot switch to a virtual console to stop X, nor shut the machine down. In fact I am lucky if I get it shutdown when I first notice it happening as before the shutdown procedure can complete its already usually dead.
Seeing as it is DIRECTLY related to having an active internet connection or not, I started thinking of everything the system loads automatically which requires an internet connection to function. Unfortunately I never get past ntpd/ntpdate. What's really baffling is that kill -9 'ing all of those process, in fact one time I did ps aux -c | grep less from the virtual console before even starting x, and killed almost everything not critical to the machine. Things like adjkerntz or cron. I figured the atrun or adjkerntz commands cron attempts may be causing it, but killing them should return the memory to the CPU, however this is literally as though the memory leaks out of it and cannot be restored.
Does anyone have ANY ideas as to what may be causing this, or even an educated guess at a list of possibilities. I have a feeling that it's a process creating a memory leak recursively, for a core
My System:
Laptop with about 1.5G Memory and 1.8G processor AMD Chipset
FreeBSD 9.0 Release
Xorg - > Gnome with no GDM management, however the exact thing happened before i built gnome and was still using twm with no XDM management.
/etc/rc.conf:
Code:
ifconfig_re0="DHCP"
moused_enable="YES"
powerd_enable="YES"
linux_enable="YES"
ntpdate_enable="YES"
ntpdate_hosts="pool.ntp.org"
dumpdev="no"
hald_enable="YES"
dbus_enable="YES"
apache22_enable="YES" //Recently added