I was also facing a similar issue and was wondering from where those files came under “Requires” category when they aren’t mentioned in the spec file. Now, I understood the problem. Thanks for the fix as well.

Have you found that the Areca cards are particularly picky with disk models and firmware versions? Their cards seem a lot better than the 3wares both under heavy IO and in recovery; but the 3ware cards are a lot easier to get up and running.


#!/usr/bin/perl
use strict;
use warnings;

sub REAPER
{ print "REAPER\n"; $RV=$?; waitpid(-1 , WNOHANG); }
$SIG{CHLD} = \&REAPER;
system('true');
print "RV: $?\n";
print "RV: $RV\n";


Since this $SIG{CHLD} = ‘IGNORE’ or \&REAPER puts -1 in $?, how do you determine if all the socket reads, writes and system calls are successful or not?

Currently one of the boxes that runs a Xen kernel is exhibiting odd behavior, seemingly random resets, that we are currently investigating; the other box seems to be doing quite well and sees a good deal of I/O.

I have not tried migrating 3ware stained drives over to an Areca controller.

At first the solution was to switch from raid6 to raid10 and that didn’t really help that much. We have major issues with high loads due to iowait and after testing some LSI cards I finally convinced some people to test out some areca’s. We got 2 of the ARC-1222′s and from my own testing they seem good but its not my decision so I am waiting to hear from my co-workers.

I use an ARC-1280ML in my home system (20x1TB raid6) and my colo’d box at work has an ARC-1220 (8×750 GB raid5). I never have iowait issues on my dated ARC-1220 controller even when seeding a bunch of linux distro ISOs which is about as IO/heavy as things can get. All the machines are 2U super-micro systems with the 8x hot-swap bays. The slightly older systems are using 8×750 GB seagate ES.2 drives and the newer ones are using the 1TB AS series (going with AS was not a good choice IMHO).

3ware has been really kick our asses not just because of performance but so far we already lost about 1% of 3TB of data due to an fsck being ran when one of the drives was messed up (getting bad sectors and what not) but the stupid 3ware controller didn’t kick it out of the array. Another system got some error with the onboard cache sync and after rebooting the array was just gone. With a livecd and tw_cli it complained about the DCB being corrupt or something which I believe is the disk configuration block which has all the raid info in and is completely unrecoverable from on the 3ware controllers. On an areca you could re-create the array (assuming you knew the settings) and chose noinit and recover from that.

Anyway these 3wares have been leaving a bad taste in my mouth and I was curious of the experience of someone else who has used areca controllers in a large amount of machines. Have any problems with DOA or controllers dieing with the arecas?

Also I assumed it wasn’t possible but I was curious if you have tried using drives from a 3ware controller and been able to use the array on an Areca?