Topic: sme box locking up..

[John]

when transfering files from one pc on my lan to the sme box via a samba share after a couple of minutes the sme box locks up...Meaning i can no longer view the samba shares on the sme box..login in thru putty and type this command "service smb restart" sometimes will allow the shares to be seen again but most times not.I have to reboot the sme box..running sme 5.6 with all updates installed..One pc is windows xp pro and another windows machine is windows 2000..Any ideas what is causing this error?

[Kelvin]

Hi John,

By your description, it does not actually sound like the box had locked up since you can still putty in and issue commands. So, what does the messages log file say ?

Kelvin

[Jakob]

Hi

Both myself and a friend had the same problem and it was the netword card that locked up. Vi both used a 3Com and after changing it to a new (and much cheaper one) we havn't had any problems since.

Jakob

[John]

Well my problem seems to be getting worse....Thought it was do to the samba shares well it's not...This problem can and does lock up the whole sme box..To the point that i can't even log in as root..This only occurs when moving files from oine hard disk to another insdie the sme box..
sme box currently has 5 hard disks installed 2 r on promise ata/100 pci card the other drives r connected to the motherboard ide channels..Locked in the samba logs nothing shows up...All my hard disks r starting to show alot of non-continous files on startup..One disk showed 49.7% non-continous files..I assume it's cause i have to hit the reset button on the box...any ideas?

[Kelvin]

OK,

Back to basics then.

1. Are you able to remove the promise controller for testing ? If so, take out the promise card (and the 2 hdds connected to it) and see if you still get a lockup when moving files.

2. Are your IDE chipsets supported natively ? If a default install on SME on your M/B shows that DMA mode is turned off for the hard disks (eg. hdparm -d /dev/hda), there is a chance your chipset is not supported natively and could be causing problems (especially if you try and force hdparm to activate dma mode).

3.You might have an IRQ conflict. Try rearranging the cards in the box.

4. You might have faulty RAM. Try replacing the RAM

5. You might have a faulty M/B. No easy way to try this short of replacing the board.

Kelvin

[John]

Below is a copy of my boot log...After further testing it appears to be a windows problem..If i use mc in a putty window i can transfer files all day long and not have a lock up occur..I transfered over 100 gigs worth of data from one disk to another all afternoon and it never locked up..


lime4x4 kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
Nov  1 11:58:54 lime4x4 kernel: VP_IDE: VIA vt82c686b (rev 40) IDE UDMA100 controller on pci00:07.1
Nov  1 11:58:54 lime4x4 kernel:     ide0: BM-DMA at 0xc000-0xc007, BIOS settings: hda:DMA, hdb:DMA
Nov  1 11:58:54 lime4x4 kernel:     ide1: BM-DMA at 0xc008-0xc00f, BIOS settings: hdc:DMA, hdd:DMA
Nov  1 11:58:54 lime4x4 kernel: PDC20267: IDE controller on PCI bus 00 dev 78
Nov  1 11:58:54 lime4x4 kernel: PCI: Found IRQ 11 for device 00:0f.0
Nov  1 11:58:54 lime4x4 kernel: PCI: Sharing IRQ 11 with 00:0b.0
Nov  1 11:58:54 lime4x4 kernel: PDC20267: chipset revision 2
Nov  1 11:58:54 lime4x4 kernel: PDC20267: not 100%% native mode: will probe irqs later
Nov  1 11:58:54 lime4x4 kernel: PDC20267: (U)DMA Burst Bit ENABLED Primary PCI Mode Secondary PCI Mode.
Nov  1 11:58:54 lime4x4 kernel:     ide2: BM-DMA at 0xe400-0xe407, BIOS settings: hde:DMA, hdf:pio
Nov  1 11:58:54 lime4x4 kernel:     ide3: BM-DMA at 0xe408-0xe40f, BIOS settings: hdg:DMA, hdh:pio
Nov  1 11:58:54 lime4x4 kernel: hda: ST33232A, ATA DISK drive
Nov  1 11:58:54 lime4x4 kernel: hdb: Maxtor 96147H8, ATA DISK drive
Nov  1 11:58:54 lime4x4 kernel: hdc: Maxtor 6Y160P0, ATA DISK drive
Nov  1 11:58:54 lime4x4 kernel: hdd: Memorex CRW-1622, ATAPI CD/DVD-ROM drive
Nov  1 11:58:54 lime4x4 kernel: hde: Maxtor 6Y060L0, ATA DISK drive
Nov  1 11:58:54 lime4x4 kernel: hdg: WDC WD800JB-00ETA0, ATA DISK drive
Nov  1 11:58:54 lime4x4 kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Nov  1 11:58:54 lime4x4 kernel: ide1 at 0x170-0x177,0x376 on irq 15
Nov  1 11:58:54 lime4x4 kernel: ide2 at 0xd400-0xd407,0xd802 on irq 11
Nov  1 11:58:55 lime4x4 kernel: ide3 at 0xdc00-0xdc07,0xe002 on irq 11
Nov  1 11:58:55 lime4x4 kernel: hda: 6303024 sectors (3227 MB) w/128KiB Cache, CHS=781/128/63, UDMA(33)
Nov  1 11:58:55 lime4x4 kernel: hdb: 120060864 sectors (61471 MB) w/2048KiB Cache, CHS=7473/255/63, UDMA(100)
Nov  1 11:58:55 lime4x4 kernel: hdc: 320173056 sectors (163929 MB) w/7936KiB Cache, CHS=317632/255/63, UDMA(100)
Nov  1 11:58:55 lime4x4 kernel: blk: queue c0365f4c, I/O limit 4095Mb (mask 0xffffffff)
Nov  1 11:58:55 lime4x4 kernel: hde: 120103200 sectors (61493 MB) w/2048KiB Cache, CHS=119150/16/63, UDMA(100)
Nov  1 11:58:55 lime4x4 kernel: blk: queue c03662b0, I/O limit 4095Mb (mask 0xffffffff)
Nov  1 11:58:55 lime4x4 kernel: hdg: 156301488 sectors (80026 MB) w/8192KiB Cache, CHS=155061/255/63, UDMA(100)

[Kelvin]

Hi John,

>After further testing it appears to be a windows problem..If i use mc in a putty
>window

Perhaps a NIC issue as previously suggested ?

Kelvin

[John]

thought that also but i've changed the nics and if it was a nic issue would it lock up the sme box totally?

[Ray Mitchell]

You say it appears to be a Windows problem and you say one pc is windows xp pro and another windows machine is windows 2000.

Does it ony happen when using one of the Win OS's (which one) or can it happen if using either one.

How much RAM in your sme server ?
Regs
Ray

[John]

It locks up using w2k or xp pro....example it will lock up if i transfer a large number of files from hdb to hdc using the samba share under windows networking..
Sme box specs r  1 ghz celeron 512 megs of ram.It was running xp pro on that computer for a year with the same hardware it worked fine..Sme 5.6 has been installed for just over a month and it was working great till about 3 days ago when it started locking up when transfering files around..Thought it was a samba issue at first cause that was the only thing that was affected..Could'nt view the samba shares thru windows network...Could putty and log into the ftp server just fine..Then it started to hard lock the sme box..Couldn't putty into box or log in to the ftp server..sme box is used as in server/gateway mode with all updates installed..

[Ray Mitchell]

Did you make ANY changes
to the sme box (server manager or command line)
or
either of the Win workstations
a few days or a week or two ago ?
Think back carefully if there was something seemingly unrelated that could be the cause of the issues you now have.

When did you add the other hard drive, was it previously working OK doing the same type of file copying ?

What version of samba do you have on the sme box ?

Regs
Ray

[Johnson]

Hi John,

It could be a bad sector problem on your HDD, because that's exactly what happen to my SME box on the weekend it lockup, however i wasn't transferring 100G of data like you.

My suggestion to eliminate this doubt is get the OEM HDD analysis program and check for bad sector, their test is quite extensive and it really tells you the current status of your HDD.

Hope this help!

JL

[SloopJohnB]

I know you have heard this before, but I am still guessing it's something to do with the NIC's. Have you tried a different brand?? I know you said you tried 3com, but that driver might be corrupted in your box. I had similar issues in an RH9.0 box using a Realtek NIC, I tried another Realtek NIC and had the same issue, when I tried an Intel card, all was ok! Good Luck!!

SloopJohnB

[John]

Yes i tried variuos nic cards..Even tried the swapping nic cards option in the sme server console..Still same error..It' not related to just one disk..Currently have 5 hard disks in the sme box..hda,hdb,hdc,hdd,hde
The lock up will occur no matter which disk i'm moving data from  or to

[Ed Form]

John wrote:
>
> Yes i tried variuos nic cards..Even tried the swapping nic
> cards option in the sme server console..Still same error..It'
> not related to just one disk..Currently have 5 hard disks in
> the sme box..hda,hdb,hdc,hdd,hde
> The lock up will occur no matter which disk i'm moving data
> from  or to

You have tried changing the patch cables and moving the link to a different section of the wall cabling I hope?

Ed Form