Topic: scsi backup probs

[Rob Leahy]

We are running an e-smith server with 2 * 10g discs set up as a raid array and seagate scsi tape drive

backups work finefor 4-5 days in a row but sometime crashes with the following messages in the message log:Sep 10 07:00:00 e-smith e-smith[5243]: Processing event: mysql-delete-dumps
Sep 10 07:00:01 e-smith e-smith[5243]: Running event handler: /etc/e-smith/events/mysql-delete-dumps/S10mysql-delete-dumped-tables
Sep 10 07:00:01 e-smith e-smith[5248]: Processing event: mysql-dump-tables
Sep 10 07:00:01 e-smith e-smith[5248]: Running event handler: /etc/e-smith/events/mysql-dump-tables/S10mysql-dump-tables
Sep 10 07:16:33 e-smith kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000014
Sep 10 07:16:33 e-smith kernel: current->tss.cr3 = 00101000, %%cr3 = 00101000
Sep 10 07:16:33 e-smith kernel: *pde = 00000000
Sep 10 07:16:33 e-smith kernel: Oops: 0000
Sep 10 07:16:33 e-smith kernel: CPU:    0
Sep 10 07:16:33 e-smith kernel: EIP:    0010:[try_to_free_buffers+18/136]
Sep 10 07:16:33 e-smith kernel: EFLAGS: 00010203
Sep 10 07:16:33 e-smith kernel: eax: 00000000   ebx: c032d9a8   ecx: 00004371   edx: 00010000
Sep 10 07:16:33 e-smith kernel: esi: 00000000   edi: c0a335c0   ebp: c032d9a8   esp: c7e77fac
Sep 10 07:16:33 e-smith kernel: ds: 0018   es: 0018   ss: 0018
Sep 10 07:16:33 e-smith kernel: Process kswapd (pid: 5, process nr: 5, stackpage=c7e77000)
Sep 10 07:16:33 e-smith kernel: Stack: 00000030 00000e00 c011d442 c032d9a8 00000008 00000006 c01222ca 00000006
Sep 10 07:16:33 e-smith kernel:        00000030 c7e76000 c01dca0e c7e761c1 c0122383 00000030 00000f00 c7ff9fc0
Sep 10 07:16:33 e-smith kernel:        c0106000 c0108acb 00000000 00000f00 c0235fd8
Sep 10 07:16:33 e-smith kernel: Call Trace: [shrink_mmap+214/300] [do_try_to_free_pages+38/120] [tvecs+7598/14080] [kswapd+103/156] [get_options+0/116] [kernel_thread+35/48]
Sep 10 07:16:33 e-smith kernel: Code: 8b 76 14 83 78 20 00 75 06 f6 40 18 46 74 0f 6a 00 e8 70 01
Sep 10 07:17:13 e-smith kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000014
Sep 10 07:17:13 e-smith kernel: current->tss.cr3 = 00b78000, %%cr3 = 00b78000
Sep 10 07:17:13 e-smith kernel: *pde = 00000000
Sep 10 07:17:13 e-smith kernel: Oops: 0000
Sep 10 07:17:13 e-smith kernel: CPU:    0
Sep 10 07:17:13 e-smith kernel: EIP:    0010:[try_to_free_buffers+18/136]
Sep 10 07:17:13 e-smith kernel: EFLAGS: 00010203
Sep 10 07:17:13 e-smith kernel: eax: 00000000   ebx: c032d9a8   ecx: 00004371   edx: 00010000
Sep 10 07:17:13 e-smith kernel: esi: 00000000   edi: c0a335c0   ebp: c032d9a8   esp: c0fabc70
Sep 10 07:17:13 e-smith kernel: ds: 0018   es: 0018   ss: 0018
Sep 10 07:17:13 e-smith kernel: Process dump (pid: 5292, process nr: 74, stackpage=c0fab000)
Sep 10 07:17:13 e-smith kernel: Stack: 00000005 00000901 c011d442 c032d9a8 00000015 00000006 c01222ca 00000006
Sep 10 07:17:13 e-smith kernel:        00000005 00000001 00000005 c0faa000 c01223e0 00000005 00001000 00001000
Sep 10 07:17:13 e-smith kernel:        c0122b90 00000005 00001000 00001000 0000000c 00000901 c0faa000 00000004
Sep 10 07:17:13 e-smith kernel: Call Trace: [shrink_mmap+214/300] [do_try_to_free_pages+38/120] [try_to_free_pages+40/52] [__get_free_pages+104/640] [grow_buffers+60/236] [refill_freelist+10/56] [getblk+286/324]
Sep 10 07:17:13 e-smith kernel:        [block_read+705/1268] [kfree_skbmem+50/64] [__kfree_skb+161/168] [unix_stream_recvmsg+631/808] [sock_recvmsg+66/180] [unix_stream_recvmsg+0/808] [sock_read+143/152] [default_llseek+0/120]
Sep 10 07:17:13 e-smith kernel:        [md_read+65/72] [sys_read+174/196] [system_call+52/56]
Sep 10 07:17:13 e-smith kernel: Code: 8b 76 14 83 78 20 00 75 06 f6 40 18 46 74 0f 6a 00 e8 70 01
Sep 10 07:24:37 e-smith dhcpd: DHCPREQUEST for 192.168.1.68 from 08:00:00:35:14:05 via eth0
Sep 10 07:24:37 e-smith dhcpd: DHCPACK on 192.168.1.68 to 08:00:00:35:14:05 via eth0
Sep 10 07:24:53 e-smith mc: /dev/gpmctl: Connection refused
Sep 10 07:24:53 e-smith mc: /dev/gpmctl: No such file or directory
Sep 10 07:25:25 e-smith dhcpd: DHCPDISCOVER from 00:60:67:3b:54:c6 via eth0
Sep 10 07:25:26 e-smith dhcpd: DHCPOFFER on 192.168.1.65 to 00:60:67:3b:54:c6 via eth0
Sep 10 07:25:26 e-smith dhcpd: DHCPREQUEST for 192.168.1.65 from 00:60:67:3b:54:c6 via eth0
Sep 10 07:25:26 e-smith dhcpd: DHCPACK on 192.168.1.65 to 00:60:67:3b:54:c6 via eth0

Anybody got any ideas???

Cheers

Rob

[Gene Cooper]

Rob Leahy wrote:
>
> We are running an e-smith server with 2 * 10g discs set up as
> a raid array and seagate scsi tape drive
>
> backups work finefor 4-5 days in a row but sometime crashes

I don't know about the specific error messages, but SCSI and backup problems are a drag...

In general try:

1) separate the tape drive on a separate SCSI controller or channel.

2) detune the SCSI parameters for the tape drive in the SCSI controller itself.

3) swap things when possible (tape drive, tapes, etc.)

4) update the drivers for the tape drive, controller, etc.

5) lastly, PERHAPS update the tape drive firmware

6) run an erase and retension on the tape just prior to the backup

G