centos 7.0 runs correctly but then disconnects

1x common debugging pattern for this type of problem.

1) In one window run...

inotifywait -mrq --format '%e %w%f' -e ACCESS --exclude '\.(bak|sw[xp])\$' /var/log > ~/fs.log 2>&1

Select allOpen in new window

2) Now in another window run this command.

while : ; do date && sleep 1 ; done

Select allOpen in new window

3) Then wait...

At some point the machine will die.

Hopefully the while loop will show the exact time of machine death... then...

The ~/fs.log file will show last few log files touched, which will help narrow down where to look for the problem.

4) Another likely command to run will be this...

while : ; do (echo "#####" && date && ps auxwwf) >> ~/ps.log 2>&1 ; done

Select allOpen in new window

This will correlate process table entries with machine death time.

For example, if someone has made the mistake of calling out to some 3rd party service (like to a CRM to record an opt-in), if the 3rd party API is slower than your machine, or 3rd party API just goes down periodically, then your machine will start piling up processes... likely PHP processes... which never end...

Once you have 100s-1000s+ of processes that never end, machine death will occur.

5) More data to capture is swap data, which is simple to capture...

while : ; do (echo "#####" && date && top -b -n 1 && sleep1) >> ~/top.log 2>&1 ; done

Select allOpen in new window

If you see swap space usage trend up + max out your space space, this means your OOM Killer (Out Of Memory Killer) triggers, which randomly kills processes, trying to keep the machine alive.

This always fails, as some important process always seems to get killed off.

Aside: If you get completely stumped, just hire someone to help with this.

Generally a long term Linux Savant will be able to... sense... "disturbances in the Force"... so they'll be able to check many things quickly, to have a good guess about problems.

The primary starting point is... how... ssh + key repeat speed... feels upon login.

Difficult to explain... and... if you've been working on Linux for decades... various Distro versions... sensation... point to where to start debugging.

are you sure about no ILO on the server? My HP DL380 SFF G6's have ILO

@David Johnson,
It does have a place to plug ILO, but it's not plugged..

Plugged. meaning no ethernet cable connected.

It might be simpler to ask what is the system running, monitor the memory utilization, , and restart the service to release the memory.

Does the freeze up occur around the same time?

Do you have an snmp type polling data monitor to help?

There is also the possibility that the hardware is starting to give up...
Or it just needs some TLC by cleaning out dust... Reseating all memory and cards...
Old servers need a bit of work at least once in a while...

Hello,
Thank you all for your reply,
This morning, we found the bellow message error displayed
After searching on internet, i found that it's related to storage raid controller disfunctions..
For information, this server has a 10T RAW partition, that is mapped as iSCSI device in another machine, and VMware vCenter Converter, does not seem to see this partition so we can, at list P2V this machine, (actual data with thin provisioning is just about 80G). Could you please tell us how we can do it?
Thank you in advance,

Check, a similar message is seen at times when the cd/DVD rom is scanned.

A.message on the absence of media in a removable drive will not lead to a system halt.

Check to confirm to which devi e the sd 0:1:0:0 points.

As to your second question,
This isn external resource not curated by VMware, what do you mean by venter can not see it.
1) only one system should access an IScai target.
Check the server presenting this 10TB volume to see which system is authorized to access it. Commonly IP based restriction/authorization.