Re: [Bacula-users] how to debug a job
2015-01-25 09:29:43
On 1/24/2015 1:12 PM, Dmitri Maziuk wrote:
> On 1/24/2015 9:00 AM, Kern Sibbald wrote:
>
>> Bottom line: contrary to what I previously thought there *might* be some
>> benefits to turning on spooling for disk Volumes, but only if you have
>> really fast spooling disks (or SSD). This is an interesting topic, and
>> it would be nice to see some real performance measures as often what one
>> expects in performance does not correspond to what computers really deliver.
> How much faster the cache tier is is only part of the equation, but even
> looking at just that: obviously, if your storage is a tape and your
> cache is a disk. Or the storage is spinning rust and cache's an ssd.
> Well, how about a 6G sata disk cache and .5G usb-2 storage? How about a
> 10k rpm "enterprise" cache disk and 5400 rpm storage disks?
>
> That's only cache to storage, there's also clients to cache speed and
> the degree of parallelism. If you only have one client then there is of
> course no point. At some number of clients & the amount of data you
> could keep the spool "full" and have one job despooling as fast as it
> can while all jobs are spooling as fast as they can. Theoretically. --
> Or 2 jobs despooling, or 3, but then you'll bottleneck on the sata
> backplane, or total usb throughput on the mobo.
>
I would add that network bandwidth easily becomes the bottleneck with
concurrent jobs.
Another question is whether or not attribute spooling is needed. If the
DB writes are the bottleneck, then of course it is. But what if the DB
is on fast SSD? In that case direct DB writes could potentially be much
faster than attribute spooling and de-spooling. I'm not sure it is
correct to assume that attribute spooling is beneficial. That too
requires testing.
------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
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>
|
- Re: [Bacula-users] how to debug a job, (continued)
- Re: [Bacula-users] how to debug a job, Thomas Lohman
- Re: [Bacula-users] how to debug a job, Josh Fisher
- Re: [Bacula-users] how to debug a job, Radosław Korzeniewski
- Re: [Bacula-users] how to debug a job, Radosław Korzeniewski
- Re: [Bacula-users] how to debug a job, Dimitri Maziuk
- Re: [Bacula-users] how to debug a job, Radosław Korzeniewski
- Re: [Bacula-users] how to debug a job, Kern Sibbald
- Re: [Bacula-users] how to debug a job, Dmitri Maziuk
- Re: [Bacula-users] how to debug a job,
Josh Fisher <=
- Re: [Bacula-users] how to debug a job, Kern Sibbald
- Re: [Bacula-users] how to debug a job, Roberts, Ben
- Re: [Bacula-users] how to debug a job, Robert M. Candey
- Re: [Bacula-users] how to debug a job, Radosław Korzeniewski
Re: [Bacula-users] how to debug a job, Luc Van der Veken
|
|
|