10-27-2014, 01:31 PM
zma
10-27-2014, 01:32 PM
When the email server is not responding, it seems posting on TaB also needs more time.
amoners
10-27-2014, 02:28 PM
I try to login VM to check it， but after wait a while， I can not get response from VM. It seem resources of VM is nervous. So I just restart VM and login then check it. There is nothing unusual when I look at log of mail server. And now, the mail server is go back.
zma
10-27-2014, 02:36 PM
(10-27-2014 02:28 PM)amoners Wrote: [ -> ]I try to login VM to check it， but after wait a while， I can not get response from VM. It seem resources of VM is nervous. So I just restart VM and login then check it. There is nothing unusual when I look at log of mail server. And now, the mail server is go back.
Working for me now.
From my experience: when there are processes that consume so much memory that the physical memory + swap are almost used up, the server may just seems stopping responding to requests even for ssh connections.
We may consider adding a larger swap partition or some physical memory to the VM if this happens frequently.
Here is a small script I used for monitoring servers' mounting which is invoked every 30 minutes. If I do not receive email after 30 minutes, there may be something wrong happened. For the email servers, watching some commands' output like 'free -m' may be sufficient. The smtp server may also need to be changed (the one here can only be used inside HKUST campus network).
Code:
#!/usr/bin/env bash
. ~/.bashrc
report=~/nfs-report.txt
host=`hostname`
date > $report
echo "@ $host" >> $report
echo "==============" >> $report
df -hT >> $report
echo "==============" >> $report
(for i in 10.0.3.{172..183}; do echo $i; ssh $i 'df -hT'; done) >> $report
cat $report | mailx -S smtp="smtp://smtp.ust.hk" -s "D-Store NFS Report @$host" -r zma@connect.ust.hk zma@connect.ust.hk