[Bacula-users] Bacula performance, LTO-3, LTO-4, block sizes
2009-01-07 18:19:59
Hi all,
I wanted to follow up to a number of threads we've already had about
Bacula performance with LTO-3 and LTO-4 drives.
The lowest-level way to test the drive performance is to use a utility
provided by the drive manufacturer. For IBM drives, there is a utility
called ITDT.
Here are my results from ITDT:
The machine is an IBM x3650 with 2 x Xeon 5150, 8GB RAM. The controller
is a SAS1068E:
> 10:00.0 SCSI storage controller: LSI Logic / Symbios Logic SAS1068E
> PCI-Express Fusion-MPT SAS (rev 02)
> Subsystem: LSI Logic / Symbios Logic Unknown device 3100
> Flags: bus master, fast devsel, latency 0, IRQ 106
> I/O ports at 6000 [size=256]
> Memory at b7ffc000 (64-bit, non-prefetchable) [size=16K]
> Memory at b7fe0000 (64-bit, non-prefetchable) [size=64K]
> [virtual] Expansion ROM at b4000000 [disabled] [size=2M]
> Capabilities: [50] Power Management version 2
> Capabilities: [68] Express Endpoint IRQ 0
> Capabilities: [98] Message Signalled Interrupts: 64bit+ Queue=0/0
> Enable-
> Capabilities: [b0] MSI-X: Enable- Mask- TabSize=1
> Capabilities: [100] Advanced Error Reporting
It connects to three LTO-4 SAS drives (in an IBM TS3310) via a 4-to-1
multiplexer dongle. 0.5m SAS cables.
With the machine idle, tape drives idle, I load a new tape into a drive
and run itdt and the "full write" test, with "compressible" and
"incompressible" data:
> Block Size Data Size Elapsed Data Rate
>
> Serial No: (KB) (MB) Time (s) (MB/s)
>
> +---------------+ +----------+----------+----------+---------+
>
> | 1310023210 | | 512 | 83923 | 1106.17 | 75.8684 |
>
> +---------------+ | 256 | 83923 | 1122.65 | 74.7544 |
>
> | 128 | 83923 | 1865.06 | 44.9977 |
>
> C/I: Total KB: Comp.Ratio | 64 | 83923 | 3354.12 | 25.0209 |
>
> +--+----------+---------+ | 32 | 83923 | 6320.65 | 13.2776 |
>
> | C| 445732624| 93.65 | | 16 | 15669 | 2288.65 | 6.84673 |
>
> +--+----------+---------+ | 8 | | | |
>
> | 4 | | | |
>
> +---------------+ | 2 | | | |
>
> Result: | ABORTED | | 1 | | | |
>
> +---------------+ +----------+----------+----------+---------+
>
> Code: Bin Log: Text Log:
>
> +------------------+ +--------------------++--------------------+
>
> | ABORTED BY USER | | 1310023210.blz || 1310023210.txt |
>
> +------------------+ +--------------------++--------------------+
>
>
>
> [R] Return to device list
>
> <[Q] Quit | Command >
>
> Block Size Data Size Elapsed Data Rate
> Serial No: (KB) (MB) Time (s) (MB/s)
>
> +---------------+ +-----------+-----------+-----------+-----------+
>
> | 1310023210 | | 512 | 83923 | 1109.09 | 75.6683 |
>
> +---------------+ | 256 | 83923 | 1141.26 | 73.5357 |
>
> +-+ | 128 | 83923 | 1873.68 | 44.7906 |
>
> Comp./Incomp.:|I| | 64 | 83923 | 3369.68 | 24.9054 |
>
> +-+ | 32 | | | |
>
> | 16 | | | |
>
> | 8 | | | |
>
> Curr. block size: | 4 | | | |
>
> +----------+ | 2 | | | |
>
> | 32 |(KB) | 1 | | | |
>
> +----------+ +-----------+-----------+-----------+-----------+
>
>
>
> Remaining (KB): Progress: Status:
>
> +---------------+ +--------------------+ +-------------+
>
> | 82426120 | | | | FULL WRITE | [A] Abort
>
> +---------------+ +--------------------+ +-------------+
>
Doing something like "dd if=/dev/urandom of=/tmp/test_file_10G bs=1M
count=10000" then "dd if=/tmp/test_file_10G of=/dev/nst0 bs=X" gives me
almost exactly the same results.
If the formatting is all screwed up, trying viewing the message as text.
Regards,
Alex
------------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It is the best place to buy or sell services for
just about anything Open Source.
http://p.sf.net/sfu/Xq1LFB
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- [Bacula-users] Bacula performance, LTO-3, LTO-4, block sizes,
Alex Chekholko <=
|
|
|