Bryn Hughes wrote (2014/11/05):
> Finally there are some OS-level settings for the 'st' driver (I'm
> assuming you are on Linux). With my LTO3 drives I need to add this to
> the kernel command line:
Hi, Solaris 11.
> On 2014-11-05 03:48 AM, Roberts, Ben wrote:
> > Hi all,
> > I'd like to try and make some speed improvements to my Bacula setup
> > (5.2.13, Solaris11). I have data (and attribute) spooling enabled
> > using a pool of 46x 1TB directly-attached SAS disks dedicated to this
> > purpose. Data is being despooled to 2x directly-attached SAS LTO6
Hmm, I do not believe in this. I think that this is too expensive and
too inefective solution. Too many mechanical disks. Which volume
organization do you use? ZFS: zpool list -v?
46 mechanical disks can provide really big aggregate speed for many
parallel data streams, but if you have just one or two 200 MB/s read
streams and several slow write streams, it is very hard work for
mechanical disks.
As wrote Bryn, it is best to obserse (peak) drives/RAID utilisation,
which gives you best valuation, if current spooling storage is
sufficient or not.
For LTO-6 tape, there is digital speed matching / data rate matching,
which can slow down physical write speed from 160 MB/s up to
40 MB/s / 54 MB/s. The tape drive tries to learn the best speed from
data stream continuity, so even very small almost unobservable delays
can result in slow tape writes. Even if you see good average
throughput per second.
> > drives at around 100mB/sec each. I think I should be able to get
> > closer to the ~160mB/s maximum uncompressed thoughput the drives and
> > tape media support (ref:
> > http://docs.oracle.com/cd/E38452_01/en/LTO6_Vol4_E1/LTO6_Vol4_E1.pdf)
You can try to switch off tape drive compression. There is some possibility,
that write speed could increase. With tape drive compression, you have to
have bigger throughput, atleast 200 MB/s.
> > I've just done a speed test and can read from the spool array at a
> > sustained 300mB/sec even while other jobs are running, so I'm sure
> > there's no bottleneck at the disk layer. My suspicion is that the
> > bottleneck is at the application layer, probably due to the way I have
> > Bacula configured.
Don't be sure. I had similar speeds and it never worked for me. Yes,
there was much lower number of mechanical drives, but 46 have not
to be better that 4 in this case.
However in any case, observe CPU utilization too. And do not forget
to compile Bacula with compiler optimizations, e. g. gcc -O3.
And I expect that you have Spool Data = yes and you do not have
Spool Attributes = no.
> > Having read through Bareos' tuning paper
> > (http://www.bareos.org/en/Whitepapers/articles/Speed_Tuning_of_Tape_Drives.html),
> >
> > I've updated the max file size from 1->50GB which increased the
> > throughput from ~75 to ~100mB/sec. I believe I need to look at tuning
> > the block size to gain the last bit of improvement.
Count 3-5 seconds per filemark. So yes, after change from 1 GB to 50 GB
file size you could see such a big improvement, but potential of change
of this type is exhausted now.
Bacula 5.2.12, FreeBSD 9.3, LSI MegaRAID SAS 9265-8i with
4 KINGSTON SKC100 480GB SSDs in RAID0, so read/write performance
is between 1-2 GB/s, LTO-5 tape drive with 140 MB/s native
speed. As you can see, there is no problem to run over
180 MB/s even on just LTO-5 an I saw up to 250 MB/s with
highly compressable data (even with just 64 KB per transfer
on tape, which is often discussed here too, that 64 KB/t is
too low). Temporal drops to 30 MB/s are due to filemarks,
file size is 16 GB.
mfid0 sa0
time KB/t tps MB/s KB/t tps MB/s
...
2014/11/06 01:49:40 119.5 2685 313.3 64.00 2657 166.1
2014/11/06 01:49:45 120.6 2345 276.3 64.00 2853 178.3
2014/11/06 01:49:50 120.3 2453 288.2 64.00 2778 173.6
2014/11/06 01:49:55 115.5 1478 166.6 64.00 602 37.62
2014/11/06 01:50:00 124.8 1419 172.8 64.00 2292 143.3
2014/11/06 01:50:05 123.5 1765 212.9 64.00 2868 179.2
2014/11/06 01:50:10 118.3 2884 333.2 64.00 2626 164.1
2014/11/06 01:50:15 121.8 1993 237.0 64.00 2331 145.7
2014/11/06 01:50:20 122.7 1979 237.1 64.00 2603 162.7
2014/11/06 01:50:25 122.6 2140 256.3 64.00 2844 177.7
2014/11/06 01:50:30 121.2 2237 264.8 64.00 2807 175.4
2014/11/06 01:50:35 122.2 2306 275.2 64.00 2973 185.8
2014/11/06 01:50:40 120.5 2711 319.0 64.00 2999 187.4
2014/11/06 01:50:45 79.67 1397 108.7 63.99 581 36.31
...
--
Rudolf Cejka <cejkar at fit.vutbr.cz> http://www.fit.vutbr.cz/~cejkar
Brno University of Technology, Faculty of Information Technology
Bozetechova 2, 612 66 Brno, Czech Republic
------------------------------------------------------------------------------
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
