Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I have one server running Debian 6.0.5 with the problem I can not fix for a few days. Everything is ok if I have running ssh console or "ping -t ". When it has been stopped everything is ok in around 10 minutes. After that we're loosing our access by web. But if I will reload the page from 4 to 7 times everything coming back. While it's happends this server can ping anything and can be pinged.

No errors or warnings in syslog, messages and apache logs. No errors in sysstat sar -n EDEV.

I've changed the network switch to exclude this from reasons.

Some files here:

/etc/network/interfaces

 #The loopback network interface
    auto lo
    iface lo inet loopback

    #The primary network interface
    allow-hotplug eth1
    iface eth1 inet static
    address 192.168.30.7
    netmask 255.255.254.0
    gateway 192.168.30.2
    network 192.168.30.0
    broadcast 192.168.31.255
    dns-nameservers 192.168.30.14 192.168.30.10
    dns-search ru

/etc/hosts

    127.0.0.1       localhost
    192.168.30.7    myhost.mydomain.ru      myhost

/etc/resolv.conf

    domain  mydomain.ru

    nameserver 192.168.30.10
    nameserver 192.168.30.14

Any ideas? Thanks a lot!

upd: no any rules in iptables. no firewall or any restrictions.

share|improve this question
2  
Could you check that there is no IP conflict on your network ? –  Dom Dec 13 '12 at 8:23
    
I've checked it first. No conflict. –  Mike Dec 13 '12 at 8:50
    
How did you check it ? Really looks like someone has the same ip... Make tcpdump -i $iterface port $wEB-port -s65535 -w network.pcap and look at dump. –  alterpub Dec 13 '12 at 8:53
    
I used arp-scan -l -g. But I tried tcpdump too and can not find anything strange there. –  Mike Dec 13 '12 at 9:45
    
Ok. I've changed an IP to another. No effect. –  Mike Dec 13 '12 at 10:01

2 Answers 2

I have seen the same issue few times in our servers. The issue we figured is that the RAM in the machine gets completely filled and no RAM left in the server, and the server gets hanged and not responding, but still able to ping.

Once few processes release some RAM then everything gets back to normal. Sometimes we had to hard reboot the machine to get it back in functional state.

To check the same, you should start monitoring your Machine with sar, that will give you some idea about the RAM usage and the trend and you would be able to get a proof whether it's really the issue.

Alternatively, you can also have top command running in one console, so that you could get an idea of the resources getting used when the server hangs. That will be a quick check but you need to be in front of the console when it happens.

share|improve this answer
    
sar -r said 46% RAM maximum used all the time. –  Mike Dec 13 '12 at 10:01
    
I don't believe that sar -r would provide you with the result of all time data, but gives you more of last few hours only. –  GeekRide Dec 13 '12 at 10:05
    
I've added to cron cat /proc/meminfo | grep MemFree >> templog.log not over 50% of used at the moment of the event. –  Mike Dec 13 '12 at 10:16
    
Yep .. that's a good idea ... run this cron every minute ... then you will get a good idea of your memory usage. –  GeekRide Dec 13 '12 at 10:17
    
Though I will suggest to change it like this ... cat /proc/meminfo | grep MemFree >> /tmp/templog.log –  GeekRide Dec 13 '12 at 10:18
up vote 2 down vote accepted

Thanks Dom and GeekRide for the good advices. Dom was right. But full picture was a little bit strange. I found an old server machine with FreeBSD. It has the same IP address and this machine was in sleep mode. But every 15 minutes it wakes up and tried to do something. The legaсy of old admin.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.