inz.fi

Unnamed repository; edit this file 'description' to name the repository.
Log | Files | Refs

commit df2af57b0881fb74e47bfbf18e5dacf3bb3f97aa
parent 5ba1406ec7013e88ed4edc600dc47657d1d25cfc
Author: Santtu Lakkala <inz@inz.fi>
Date:   Tue, 25 Jul 2006 10:05:52 +0000

Upgrading the router to file server

Diffstat:
Aposts/upgrading-the-router-to-file-server.md | 24++++++++++++++++++++++++
1 file changed, 24 insertions(+), 0 deletions(-)

diff --git a/posts/upgrading-the-router-to-file-server.md b/posts/upgrading-the-router-to-file-server.md @@ -0,0 +1,23 @@ +# Upgrading the router to file server + +Started upgrading the router machine to be a file server too. Got two Delock PCI SATA controllers (SiI3112A chipset) and four Western Digital 320 GB platters, with intention to have ~1 TB RAID5 on the machine. + +The problems started before even physically connecting the disks, or the controllers. The casing only has places for 4 3,5" devices, and I intend to keep floppy drive and old ATA -disk there too. Now two disks are "lying" on the bottom of the case (put some legs on them, so they're not exactly lying there). Problem solved 'nuff (maybe I'll buy some 5,25"/3,5" adapters later). + +The problems didn't stop, quite the opposite actually. While adding the cards and disks physically was quite a breeze, (added also sound card and replaced realtek based card with 3com one, all pci slots full now) the machine refused to start debian from the old disk. In fact, it never got to LILO. + +Then I tried to disconnect the SATA drives from their power source, and we got a running debian with no sata disks. Some fiddling, like installing new kernel, and shutdown; this time startup to LILO with drives unpowered, powering them up and continued loading Linux with the drives. This worked suprisingly well, except that the drives on the latter controller gave some errors. + +Now I've got the system up, but somewhy bringing eth1 up (pcnet32/HomePNA) freezes the whole system, eth0 (3c59x) works just fine. I still need to figure out this problem in order to get the network running again and also find out why the 2nd controller fails. It may have something to do with them seemingly being on the same IRQ, which means some BIOS/CMOS hacking. + +The battle continues today, I ain't gonna lose. =) + +**Update:** + +It works, yay: +<pre> +inzane@h0x:~$ df -h /mnt/tmp/ +Filesystem Size Used Avail Use% Mounted on +/dev/md0 881G 33M 836G 1% /mnt/tmp +</pre> +And all I had to do was to remove SB Live! and put the HomePNA adapter in its steadt and tweak the IRQ's in CMOS (all my PCI cards were on the same IRQ - not good). Not sure though, if I could boot the machine without hot-powering the SATA drives. =) +\ No newline at end of file