[BlueOnyx:11521] Re: Aventurin problem

[Steffan]

The machine is 2 1/2 years old
So it  is still in his guarantee

We are going to try to replace the areca raid card
On google more peaple had the same problem
Hoping that fixe it

Also im going to migrate all servers of this one in a while and do a
reinstall.


-----Oorspronkelijk bericht-----
Van: blueonyx-bounces at mail.blueonyx.it
[mailto:blueonyx-bounces at mail.blueonyx.it] Namens George F. Nemeyer
Verzonden: vrijdag 12 oktober 2012 2:08
Aan: BlueOnyx General Mailing List
Onderwerp: [BlueOnyx:11519] Re: Aventurin problem

On Thu, 11 Oct 2012, Steffan wrote:

> Hope someone or Michael can help
> My aventurin server is starting to crash

How old is the machine?

Usually, when machines start randomly acting flakey, the filter capacitors
on the motherboard and/or power supply are starting to fail.  Age and heat
will dry out the electrolyte, and they basically begin letting the power
supply lines become 'noisy'.  Some of the capacitors can actually bulge or
burst and spew electolyte from excess internal heat build up.  There's even
been a few major scandals over some brands being prone to fail in past
years.

Symptoms can be locking up, strange error messages, and corrupted data, and
other, often intermittant bad behavior.  Locking up is the most common,
since any messed up memory reads frequently leads to programs or the OS
crashing when the CPU heads off into the weeds executing bad instructions.

Generally, things will worsten until the machine won't even boot completely.

Bad/failing memory (or CPU) can also produce similar symptoms.

> It could be a hardware problem but i also notices after the reboot the 
> server spitssout this messages
> kernel: Route hash chain too long!
> kernel: Adjust your secret_interval!

It could have garbaged the file that's being complained about.  Noise can
flip bits on data lines, so random sectors getting written to wrong places
can happen.

> Can this be related ?
>
> 2.6.18-194.26.1.el5.028stab079ave02.2ent on a i686

Also, your comment about the system bootup fixing 'orphaned inodes' is
typical if the machine locks up in the middle of disk writes, leaving the
low level sector indexing in some strange state.  The OS tries to check and
fix what it can if it sees the machine wasn't shut down cleanly.

You can try replacing the memory to see if that helps.  Sometimes just
unplugging and reseating it may be the problem.  Likewise for various
plug-in cables.  Also, make sure the fans and any air filters are ok, since
overheating can cause crashing as well as shortening the life of various
components.

If the box is several years old, it may be time to retire it, or if it's
under a warranty, having it replaced.

=^_^=  Tigerwolf
_______________________________________________
Blueonyx mailing list
Blueonyx at mail.blueonyx.it
http://mail.blueonyx.it/mailman/listinfo/blueonyx