Problems caused by adding new drives to my file-server

[klart]

Hi!

I have a file-server with 14 HDD for the moment(10 IDE + 4 SATA). The problem is, that when I add a new ard drive, the system(Fedora 4) goes a bit crazy. The IDE drives cause some errors in the kernel and get remounted in read-only mode.
This problem only appears when I'm adding a new drive, if I reinstall the system with all the drives attached, everything is fine again. All the drives are independent, no LVM or RAID.

I'm not sure if anyone might know what the problem is, but I gave it a try.

Thank's in advance!

Specs.:
AOpen AK77
Promise 2 channel IDE-RAID PCI
10 IDE Maxtor HDD
4 SATA Maxtor HDD

 

[hulksterjoe]

you mean aside from the fact that ide drives arent hotswapable.. i'd imagine thats why the kernel gets screwed up and shy it works on the reboot

 

[klart]

If the solution would be so simple...
No, of corse, I'm tolking about a normal HDD addition, with the computer turned off. And when it gets OK again is the time when I reinstall the system. Not reboot but reinstall.
It's probably a lot more of a linux question... I just tought that those, who are interested in storage systems could have already met a problem like this...

 

[unhappy_mage]

Check your /etc/fstab; if there are different partitions on the drive that get swapped out versus the one that comes in, there could be problems. What error exactly is it? Try "dmesg" to show kernel errors, and "mount -a" will probably show whatever complaints come up outside the kernel.

And what PSU do you have running all those?

 

[kuyaglen]

And what PSU do you have running all those?

I would like to know this too...I currently have an Allied 500W powering my fileserver but
I'm gonna add to more but I'm sure I'm gonna need at least a 550W...

P4 3.06 Northwood
Intel D865GBF mobo w/onboard VGA
1024 Corsair Value ram (2x512)
Airlink+ gig nic
Promise SX4000lite raid controller (w 256mb pc133)
Syban (Fry's cheapest) IDE/raid controller

(currently 2.8TB)

yet to install
2nd Syban ide/raid controller
2 more Maxtor DiamondMax 10 HDDs

 

[klart]

My PSU is a Héden 560W power supply. It's not a big brand, it's sold by LG France. But according to the BIOS voltage monitoring, everything is fine.
I'm not replacing the drives but I'm adding new ones. So there is no confusion caused by the pratitions type or number. All of the drives has a single EXT3 partition except for the system drive, which has a SWAP too. All of these partitions are primary ones.

Oct 24 06:35:18 fserv1 kernel: hdd: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Oct 24 06:35:18 fserv1 kernel: hdd: dma_intr: error=0x40 { UncorrectableError }, LBAsect=208142399, high=12, low=6815807, sector=208142399
Oct 24 06:35:18 fserv1 kernel: ide: failed opcode was: unknown
Oct 24 06:35:18 fserv1 kernel: end_request: I/O error, dev hdd, sector 208142399
Oct 24 06:35:18 fserv1 kernel: EXT3-fs error (device hdd1): read_block_bitmap: Cannot read block bitmap - block_group = 794, block_bitmap = 26017792
Oct 24 06:35:18 fserv1 kernel: Aborting journal on device hdd1.
Oct 24 06:35:18 fserv1 kernel: ext3_abort called.
Oct 24 06:35:18 fserv1 kernel: EXT3-fs error (device hdd1): ext3_journal_start_sb: Detected aborted journal
Oct 24 06:35:18 fserv1 kernel: Remounting filesystem read-only
Oct 24 06:35:18 fserv1 kernel: EXT3-fs error (device hdd1) in ext3_prepare_write: IO failure

That's the kind of error I get. It only appears when I try to write on a drive, reading is always fine.