On Fri, 10 Nov 2006, Rudolf Cejka wrote:
> Alan Brown wrote (2006/11/10):
I did not do /dev/zero because of hardware compression. I guess I could
>>> turn that off but isn't /dev/random 1 byte at a time??
>> It's fast enough on most systems.
>
> Really? My experience is that the speed of /dev
Alan Brown wrote (2006/11/10):
> >> I did not do /dev/zero because of hardware compression. I guess I could
> > turn that off but isn't /dev/random 1 byte at a time??
> It's fast enough on most systems.
Really? My experience is that the speed of /dev/random and /dev/urandom
in FreeBSD is far from
On Thu, 9 Nov 2006, John Drescher wrote:
>> > Have you tried a test of the raw tape speed? When I first got mt lto-2
>> > library I did a few tests like the following
>> >
>> > time dd if=/dev/sda of=/dev/nst0 bs=1G count=10
>> >
>> > and I got around 35MB/s.
>>
>> Try if=/dev/zero and if=/dev/ra
On 11/9/06, John Drescher <[EMAIL PROTECTED]> wrote:
> > Try if=/dev/zero and if=/dev/random
> >
> > AB
Note that depending on the platform where you try,
/dev/urandom as a source might be faster than /dev/random.
("/dev/random blocks until more entropy can be obtained" says
e.g. the Solaris man p
As a reminder, the director/sd/database system is supporting 20
concurrent jobs, 10 to tape and 10 to raid. It's a 2-cpu hp dl380 G4,
3.4GHz cpus and 3.5Gb ram.
On Thu, 2006-11-09 at 17:29 -0500, John Drescher wrote:
> On 11/9/06, Don MacArthur <[EMAIL PROTECTED]> wrote:
> The jobs are
On 11/9/06, Don MacArthur <[EMAIL PROTECTED]> wrote:
The jobs are written to both the raid and the tape at almost identicalspeeds. I don't know what my total throughput is,but the transfer rate
for each job is about 1.5mbps to 3mbps.For a few minutes after reading this I was very confused about th
The jobs are written to both the raid and the tape at almost identical
speeds. I don't know what my total throughput is, but the transfer rate
for each job is about 1.5mbps to 3mbps.
On Thu, 2006-11-09 at 16:44 -0500, John Drescher wrote:
>
> On 11/9/06, Don MacArthur <[EMAIL PROTECTED]> wrote
On 11/9/06, Don MacArthur <[EMAIL PROTECTED]> wrote:
I have a 7-disk raid fibre attached and a lto3 scsi attached to my sd,and I get the same throughput for both. I run 10 jobs (two from eachclient) to each store concurrently.
Both what? What backup speeds are you getting? Thanks,John
---
I have a 7-disk raid fibre attached and a lto3 scsi attached to my sd,
and I get the same throughput for both. I run 10 jobs (two from each
client) to each store concurrently.
I don't think the medium or the connection are the holdup. My network
throughput at that time varies around 750 mbps.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Adam Huffman wrote:
>> Is the data compressable?
>>
> A mixture - some already compressed, some images, some XML data etc.
>
> The library and the disks are on the same fibre channel fabric, so I
> wouldn't expect that to be a bottleneck.
>
> I'm us
On 11/9/06, Alan Brown <[EMAIL PROTECTED]> wrote:
On Thu, 9 Nov 2006, John Drescher wrote:> Have you tried a test of the raw tape speed? When I first got mt lto-2> library I did a few tests like the following>> time dd if=/dev/sda of=/dev/nst0 bs=1G count=10
>> and I got around 35MB/s.Try if=/dev/z
> From: "Adam Huffman" <[EMAIL PROTECTED]>
>
> What sort of backup speed should I expect when using Bacula with an
> LTO3 library?
>
> I'm seeing nothing like the advertised 60-80MB/s.
The tape drive speed is not the bottleneck, then.
The number of files saved (or, seen from another point, the
On Thu, 9 Nov 2006, John Drescher wrote:
> Have you tried a test of the raw tape speed? When I first got mt lto-2
> library I did a few tests like the following
>
> time dd if=/dev/sda of=/dev/nst0 bs=1G count=10
>
> and I got around 35MB/s.
Try if=/dev/zero and if=/dev/random
AB
-
On 11/9/06, Carlos Cristóbal Sabroe Yde <[EMAIL PROTECTED]> wrote:
The top speed I've achieved with an IBM 3581 L38 was around 30 MB/s on a test(btest).Have you tried a test of the raw tape speed? When I first got mt lto-2 library I did a few tests like the following
time dd if=/dev/sda of=/dev/nst
The top speed I've achieved with an IBM 3581 L38 was around 30 MB/s on a test
(btest).
With bacula, local disk (sata):
JobId: 1384
Job:srv-backup.2006-10-27_22.05.07
Backup Level: Full
Client: "srv-backup" x86_64-unknown-linux
Adam Huffman wrote:
> What sort of backup speed should I expect when using Bacula with an
> LTO3 library?
>
> I'm seeing nothing like the advertised 60-80MB/s.
>
>
> Adam
>
> -
> Using Tomcat but need to do more? Need to suppor
On 08/11/06, Alan Brown <[EMAIL PROTECTED]> wrote:
> On Wed, 8 Nov 2006, Adam Huffman wrote:
>
> > What sort of backup speed should I expect when using Bacula with an
> > LTO3 library?
> >
> > I'm seeing nothing like the advertised 60-80MB/s.
>
> What speeds _are_ you seeing?
>
Around the 30MB/s m
On Wed, 8 Nov 2006, Adam Huffman wrote:
> What sort of backup speed should I expect when using Bacula with an
> LTO3 library?
>
> I'm seeing nothing like the advertised 60-80MB/s.
What speeds _are_ you seeing?
Is the data compressable?
--
What sort of backup speed should I expect when using Bacula with an
LTO3 library?
I'm seeing nothing like the advertised 60-80MB/s.
Adam
-
Using Tomcat but need to do more? Need to support web services, security?
Get stuff
19 matches
Mail list logo
