So my Raq3 with Raq4 OS does freeze regulary, or to be precise, it's immeadialy blocking traffic out of the blue sometimes. From the logs I see that the server actually is alway still running. It's just blocking all in and outgoing traffic for some reasons. It's happening very spradicly, I already had 50 days without any problem, and than again it's happening nearly daily. No unusual entries in the logs at all. What is really strange about the problem is that even a soft reboot isn't solving the problem, just a hard rreboot by powering off the box.
Any ideas?

Sorry for the late reply, but how are the fans running? What's the temperature of the CPU?
Are any drive partitions full?

Originally posted by IFAPaulB
Sorry for the late reply, but how are the fans running? What's the temperature of the CPU?
Are any drive partitions full?

Fans are running fine, and actually just where renewed some months ago. Temperature is never above normal, I checked that.
/dev/hda1 is az 87%, but I guess that's normal.

Sounds like your power supply is borked.... I had to replace two powersupplies myself lately due to similar problem.

Apparently the powersupplies 'give up' after 4 to 5 years.

Originally posted by Floid
Fans are running fine, and actually just where renewed some months ago. Temperature is never above normal, I checked that.
/dev/hda1 is az 87%, but I guess that's normal.

But the server is still running when the problem occurrs, it's just not online anymore and it blocks all traffic for some reasons. If it would be a problem with the power supply, wouldn't the server just crash entirely?

If the powersupply would not power the server anymore, yes. But if the PSU supplies very unleveled power, this is one of the things that can happen.

You mentioned "from the logs" that the server is still running.

When a crash occurs, does the "service checker" still check Web and mail every 15 minutes? Are those log entries still made?

Check /var/log/messages for any kernal "Oops" or panic errors.

Originally posted by IFAPaulB
You mentioned "from the logs" that the server is still running.

When a crash occurs, does the "service checker" still check Web and mail every 15 minutes? Are those log entries still made?

Check /var/log/messages for any kernal "Oops" or panic errors.


Yes, those entries are still being made.

Originally posted by netvictory
If the powersupply would not power the server anymore, yes. But if the PSU supplies very unleveled power, this is one of the things that can happen.

That's a possibility. I've never had a power supply intermittently glitch like that...the only one's I've had fail fail completely.

However, most Raq4's are about 4+ years old now, after running 24/7. I guess anything's possible :stickout: Kind of like a car with 300,000 miles on it.

Originally posted by netvictory
If the powersupply would not power the server anymore, yes. But if the PSU supplies very unleveled power, this is one of the things that can happen.

How likely is that? Could you give me any percentage number? Would you actually advice me to buy a new power supply?

Computer power supplies usually have some compensation circuitry built in so when the input voltage fluctuates, the output stays constant. Computers are VERY sensitive to voltage fluctuations-even half a volt can really cause things to get out of whack.

I would check the things I mentioned above, and check software related issues before spending money on new hardware.

Originally posted by IFAPaulB
Computer power supplies usually have some compensation circuitry built in so when the input voltage fluctuates, the output stays constant. Computers are VERY sensitive to voltage fluctuations-even half a volt can really cause things to get out of whack.

I would check the things I mentioned above, and check software related issues before spending money on new hardware.

Like I told you, i already checked that possibilies, and I can rule them out.
If it would really be a problem with the power supply, which should always only the (onboard) NIC be effected, so constantly? I mean I never have any other hardware problems, if it really would be a power supply problem, wouldn't things like the hd be afftected first?

We had this issue where the NIC kept dropping out. I made a change to the modules.conf file to add some network parameters. I think we simple force the nic into the correct mode (10/100 half/full). I will try to find my notes.

Our symptons were that the server was up and running but the network would die out. Rebooting always fixed the server.

You may need to adjust this for your network but you may want to add:

options eepro100 options=0x30,0x30

to /etc/modules.conf

I think this is 10Mbit full duplex. I know that the epro100 driver was updatd in one of the kernel updates and the specs changed.

Personally I have such an incident that is happening. The whole LCD will just blank out and then nothing works.

I changed the motherboard and that stopped.

Why not, if you have another unit try putting the HDD there and see if it jams again.