Good day,
My first in your situation is "Do I have enough resources to run this
configuration?"
Check if you have enough HBA's to keep the 12 LTO3's. A single 4GB HBA
will probably not keep a single LTO3 engaged fully as that HBA can only
push a maximum of 444 MBs per second.
>>From what I can tell you do not have enough CPU, memory or HBA's to keep
those 12 tape drives busy.
The fact that your system is telling that it is waiting for resources
tells me exactly that.
Most planners I have dealt with have always made this error. They always
seem to add enough tapes drives to theoretically move that data within
the backup window, but never contact the Tape drive manufacturers as to
the requirements of the drives. I would suggest adding more HBA's,
Memory and Processors. This will allow the data to flow more freely.
Remember you can open the tap as much as you like but you will only get
out as much water as the thickness of the pipe will allow.
Netbackup have a very basic requirements document, giving you the
minimum CPU, Memory and HPA requirements.
Obey these rules and you will have no speed problems. There after it
will be tuning the backups and you will be able to utilise your
resources fully and actually finish your backups within your prescribed
window.
This is not an easy sell, backups and recovery does not come cheep. I
have had changes made in a large company and have been able to prove
this.
Good Luck.
Kind Regards,
Clem Kruger
-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of tzurita
Sent: 09 August 2007 02:01 AM
To: VERITAS-BU AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Netbackup 6.0 Backup Speed on HP Windows Blade
All, have a question about speed.
We are running the master server on a Windows C class HP blade with a
dual core 2.4 GHz Atholon CPU and 8 GBs memory. Has two FC ports
connected to:
2 Two HP StorageWorks Enterprise Class ESL712 Tape Libraries
712 - Tape slots each library
12 - LTO3 Tape drives total
Backups are not that fast and I am questioning why. This is just a fast
POC test but I would like to tune and give the customer more valid
numbers.
Any ideas would be great. My thoughts are that we don't have enough
CPU's to push the tape libraries but why such a vast difference betwen
windows and solaris backups? I am just puzzled. Help please.
Here are the values in the following files:
C:\Program Files\Veritas\Netbackup\db\config\NUMBER_DATA_BUFFERS
132
C:\Program Files\Veritas\Netbackup\db\config\NUMBER_DATA_BUFFERS_DISK
32
C:\Program Files\Veritas\Netbackup\db\config\NUMBER_DATA_BUFFERS_RESTORE
64
C:\Program Files\Veritas\Netbackup\db\config\SIZE_DATA_BUFFER
262144
C:\Program Files\Veritas\Netbackup\NET_BUFFER_SZ
1049600
C:\Program Files\Veritas\Netbackup\NET_BUFFER_SZ_REST
525312
For Network Windows servers clients, we average just under 7 MB/sec
Same server configuration for Solaris 33 MB/sec
>>From bptm log:
15:08:25.456 [3492.6020] <2> io_set_sendbuf: setting send network buffer
to
525312 bytes
15:08:25.456 [3492.6020] <2> io_init: using 128 data buffers
15:08:25.456 [3492.6020] <2> io_init: buffer size for read is 262144
15:08:25.456 [3492.6020] <2> io_init: child delay = 20, parent delay =
30
(milliseconds)
15:08:25.456 [3492.6020] <2> create_shared_memory: shm_size = 33557564,
buffer address = 0x1170000, buf control = 0x3170000, ready ptr =
0x3170c00, res_cntl = 0x3170c04
15:08:25.456 [3492.6020] <2> setup_bpbkar_info: Global\NetBackup Media
Manager SHM Info Path app429d_1186419216 file successfully created
15:08:25.456 [3492.6020] <2> nbjm_media_request: Passing job control to
NBJM, type READ
.......
15:17:12.749 [3664.1404] <2> bptm: EXITING with status 0 <----------
15:18:25.355 [2720.5456] <2> write_data: writing block shorter than
BUFF_SIZE, 218112 bytes
15:18:25.355 [2720.5456] <2> write_data: writing short block, 218112
bytes, remainder 0
15:18:25.355 [2720.5456] <2> write_data: waited for full buffer 63961
times, delayed 80764 times
15:18:25.355 [2720.5456] <2> write_backup: write_data() returned,
exit_status = 0, CINDEX = 0, TWIN_INDEX = 0, backup_status = 0
15:18:25.355 [2720.5456] <2> io_terminate_tape: writing empty backup
header, drive index 0, copy 1
15:18:25.355 [2720.5456] <2> io_ioctl: command (0)MTWEOF 1 from
(bptm.c.8229) on drive index 0
15:18:31.933 [2720.5456] <2> io_terminate_tape: absolute block position
prior to writing empty header is 241703, copy 1
15:18:31.933 [2720.5456] <2> io_write_back_header: drive index 0,
empty_file, file num = 2, mpx_headers = 0, copy 1
15:18:31.933 [2720.5456] <2> io_close: closing C:\Program
Files\VERITAS\NetBackup\db\media\tpreq\drive_HP.ULTRIUM3-SCSI.000, from
bptm.c.8369
15:18:31.933 [2720.5456] <2> io_terminate_tape: block position check:
actual 241703, expected 241703
15:18:31.933 [2720.5456] <2> send_MDS_msg: MEDIADB 1 28 AK6501 4000002
*NULL* 21 1186511708 1186511708 1187721308 0 0 0 0 1 1 0 0 1024 0 241703
0
15:18:31.933 [2720.5456] <2> ParseConfigExA: Unknown configuration
option on line 56: RenameIfExists = 1
15:18:31.949 [2720.5456] <2> JobInst::sendIrmMsg: returning
15:18:31.949 [2720.5456] <2> openTpreqFile: tpreq_file: C:\Program
Files\VERITAS\NetBackup\db\media\tpreq\drive_HP.ULTRIUM3-SCSI.000,
serial_num: HU10709CR7
15:18:31.949 [2720.5456] <2> get_drive_path: SCSI coordinates {0,0,0,1},
dos_path \\.\Tape0, pnp_path
\\?\scsi#sequential&ven_hp&prod_ultrium_3-scsi&rev_l63w#5&23bbaf37&0&000
001#{53f5630b-b6bf-11d0-94f2-00a0c91efb8b}
15:18:31.949 [2720.5456] <2> check_serial_num: serial number match for
drive with SCSI coordinates {0,0,0,1}, dos_path \\.\Tape0, drive serial
number HU10709CR7, expected serial number HU10709CR7 15:18:31.980
[2720.5456] <2> init_tape: \\.\Tape0 (SCSI coordinates
{0,0,0,1}) configured with blocksize 0
15:18:31.996 [2720.5456] <2> init_tape: \\.\Tape0 (SCSI coordinates
{0,0,0,1}) has compression enabled
15:18:32.011 [2720.5456] <2> io_open: SCSI RESERVE
>>From status log of NB Master Full backup: about 22 MB/sec
8/7/2007 2:35:56 PM - positioning AK6501 to file 1
8/7/2007 2:36:06 PM - positioned AK6501; position time: 00:00:10
8/7/2007 2:36:06 PM - begin writing
8/7/2007 2:45:18 PM - Warning bpbrm(pid=3900) from client bmedia403: WRN
- Removable Storage Management: unable to export database (WIN32 21: The
device is not ready. )
+----------------------------------------------------------------------
|This was sent by tzurita AT gmail DOT com via Backup Central.
|Forward SPAM to abuse AT backupcentral DOT com.
+----------------------------------------------------------------------
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|
