Debian crashes to a black screen with a static text cursor

[eludema]

Hi, folks!

Linux semi-newbie here. I've managed to put together a cheap web-server for the in-laws, and have done the best I can to lock it down in every way I can handle via online recommendations and tutorials. Everything seems to be working nice and smooth - until it doesn't!

The general flow is this: Boot up the machine. Make sure things are locked down (iptables blocks pretty much everything on start up). The apache server will be functional for a few days. Suddenly, things will start to not work on the site. When I try to run many commands, either as a local user or with sudo, I start getting "I/O Errors"... something that always makes me nervous. Looking at the logs and running disk checks, things seem to be fine. I finally got to visit the in-laws and see what happens to the actual server locally (before, I've always had to SSH in). After things start to not work, things become completely non-functional. Looking at the monitor hooked up to the server, the nice GUI login screen is gone and all that's left is a black "text-mode" screen with a static, non-blinking cursor in the top-left corner. All I can seem to do is a hard-reboot. Once it's rebooted, it's fine again - until it's not!

I'm using a cheap Intel Atom based server I tossed together myself a few years back. Only 2GB of Ram with a SSD drive. After not having found a smoking gun in the logs, I'm wondering if the thing is just overheating (which would be pretty pathetic considering just how little the thing is used...)

So, for the Linux semi-newbie, (before sending a dump of every log I can find), what other logs should I look at, what hardware specs, and what other obvious things should I be looking at to give me some hints as to why this things face plants after a few days of relatively light use?

Thanks, and Noob-Flame away!

E

[dasein]

...Noob-Flame away!

You got it!

First off, if this is a public-facing server, do NOT run a GUI. Period.

Second, you need to provide actual details, not your pre-digested, subjective impressions. That means detailed hardware and software specs, verbatim error messages, specific steps you've undertaken, etc. (Heavy on the "etc." part.)

If it sounds like I'm basically asking you to rewrite your entire post, then you're hearing correctly.

That said...

- Do NOT "hard reboot." Learn about the "magic" sysreq keys.

- Unless you've got zero airflow, heat is probably not the issue. (A halfway-decently-ventilated Atom doesn't run warn to the touch.) But a basic blow-out-and-reseat certainly won't hurt anything. Recommended.

- Unless your in-laws are trying to run a hosting service, your hardware seems more than adequate for lightweight httpd duties. The Web's earliest servers ran on FAR less.

[kedaha]

I start getting "I/O Errors"...

I think that an SSD card is not the best option for a web server and it may well be that it is failing owing to high write cycles to it but as dasein says, some detailed hardware and software specs are needed.

[GarryRicketson]

The linux command 'dmesg' is good for getting some info (sometimes) :

Code: Select all

man dmesg

DMESG(8) System Manager's Manual DMESG(8)

NAME
dmesg - display the system message buffer

SYNOPSIS
dmesg [-s] [-M core] [-N system]

DESCRIPTION
dmesg displays the contents of the system message buffer. It is most
commonly used to review system startup messages. On some systems the
message buffer can survive reboot and be retained (in the hope of
exposing information from a crash).

The options are as follows:--------snip-----

Also useing keywords :

Code: Select all

How to troubleshoot a linux server 

A search will reveal some tutorials that should be of help with learning how to
trouble shoot the system.

I don't know how accurate this is , but it was in the results of a search:
Using key words:

Code: Select all

I start getting "I/O Errors"

 

Solid State Drives:

SSDs are the most promising storage drive models that offer high-performance operations, but, unfortunately, do not provide a reliable durability under even the simplest of faults like power loss. Solid state drives are not that solid and comparatively sensitive than magnetic drives. The main cause of I/O errors on a solid state drive is power outages. According to major research, there is no real way to fix I/O device errors on solid state external hard drives.

At this point we do not even know if the OP is using Debian, and which version,
is the server NGINX or what ?
What is thIs ?

the nice GUI login screen

I mean is a a Gnome, login screen, or ?????
Could be Ajenti, or is it Cpanel ?
0 details = 0 answers

[acewiza]

My WAG is some power-related SSD issue. An adequate UPS for that system would be fairly inexpensive.

A more robust build might include a conventional hard drive for the web data and a UPS. Still, not really alot of cash for a server.