comparison with rival projects
A message that is not in the archive (sent off-list or not yet received): 824216649.28481292460292573.JavaMail.root@shiva
First of all, I would like to thanks for answers.
One more question and some reply in text.
After making ssd disc a cache for some partition, there is possibility to detach cache. What should I do to "erase" (trim) whole disk for another benchmarking? I think that "dd if=/dev/zero of=/dev/sdx" wouldnt be right command.
Quoted 76 lines · ----- "Nauman Rafique
Quoted 4 lines · Kent Overstreet
Quoted 3 lines · On 12/15/2010 03:26 PM, [email protected] wrote:
Hi,
Iam new in problematic of caching persistent data. I want to usecaching
Quoted 1 line
Quoted 1 line
(on ssd disk) on my desktop with a lot of ram. I hope, it will
provide me
Quoted 1 line
Quoted 1 line
better latency for starting programs (developer/scientific
desktop). My
Quoted 3 lines
Quoted 3 lines
questions are:
Is it possible to use single ssd disk for making two caches? One
for
Quoted 1 line
Quoted 1 line
/dev/md2(root partition, raid10) and second for/dev/md3(raid5,home
Quoted 1 line
Quoted 1 line
partition, etc) (I hope this is possible by partitioning that ssd
disk and
Quoted 3 lines
Quoted 1 line
using partitions for caching)
Yep. One cache device can be used with (currently) up to 256
backing
Quoted 4 lines
devices.
Quoted 2 lines
Is it dangerous (in terms of data loss) to use writethrough caching
to
Quoted 1 line
Quoted 1 line
single ssd disk of raid5 block device? (I have 5 disks, using lvm
on top of
Quoted 3 lines
Quoted 1 line
raid).
Shouldn't be, recovery from unclean shutdown is quite well tested,
both
Quoted 4 lines
writethrough and writeback should be perfectly safe.
Quoted 2 lines
For caching device is most critical number for random reads or
operations
Quoted 3 lines
Quoted 1 line
per second. Is it right?
For writethrough caching random reads, for writeback both reads and
writes.
Quoted 3 lines
Quoted 3 lines
Trim is not need feature for cache ssd device. It will have same
cache
Quoted 3 lines
Quoted 1 line
performance without it. Is it right?
Really just depends on the ssd; some (cheaper, earlier) drives are
known for
Quoted 1 line
performance significantly degrading over time, with trim helps with
(but
Quoted 4 lines
doesn't necessarily eliminate). Bcache uses trim if it's available.
Quoted 2 lines
What benefits/disadvantages have bcache [1] project over dm-cache
[2] and
Quoted 1 line
Quoted 1 line
flashcache [3]? My current knowledge is that writethrough is
possible only
Quoted 1 line
Quoted 1 line
in bcache, but dm-cache is implemented in more "standard" way
(thdough
Quoted 9 lines
Quoted 5 lines
devicemaper). Also bcache has some more information on web.
[1] - http://bcache.evilpiepirate.org/
[2] - http://users.cis.fiu.edu/~zhaom/dmcache/index.html
[3] - https://github.com/facebook/flashcacheFlashcache is based off of dm-cache.
Flashcache has been used in production awhile, bcache is still a
little
Quoted 1 line
rough around the edges - but bcache has better performance, more
features,
Quoted 1 line
and it always orders writes correctly so as to be crash safe
(flashcache has
Quoted 8 lines
Quoted 1 line
a "torn write" problem).
Kent, now that we are on the topic, I wonder if there has been some benchmarking comparing bcache to flashcache performance?
Also I wonder if you plan to target for getting bcache included in mainline kernel. I wonder if such attempts were made for flashcache but that's probably off-topic.
I hope, next two links will help in making comparison.
http://www.mysqlperformanceblog.com/2010/05/10/flashcache-first-experiments/
http://www.mysqlperformanceblog.com/2010/05/18/flashcache-more-benchmarks/
There is simple methodology of cache testing by using it with mysql database. I hope, someone will catch this discussion and try some comparable test on bcache.
For me will be next test sufficient:
Activate caching and make for example 20 reboots. Measure time from computer bios beep till starting gimp, which will be in gnome/kde/fluxbox autostart script. Of course before starting gimp, there will be need for starting some other programs like firefox, eclipse, mplayer with some film...
I would like to see these times of booting in graph. I promise that after installation of bcache I will do this test with bcache and post it here.
Quoted 21 lines
Thanks.
--
NaumanQuoted 5 lines
Quoted 3 lines
PS: Is there any tutorial for gentoo users?
Nah, I'm an ubuntu/debian user. For caching / the important thing is
to hook
Quoted 1 line
into your initramfs and get everything loaded before you mount your
root
Quoted 8 lines
filesystem.
Quoted 6 lines
Thanks for any explanations / answers,
--
LuVar
--
To unsubscribe from this list: send the line "unsubscribe
Quoted 10 lines
linux-bcache" in
Quoted 5 lines
Quoted 2 lines
the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribelinux-bcache" in
Quoted 3 lines
the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
PS: I am now decided to use bcache (after buying some ssd).
--
To unsubscribe from this list: send the line "unsubscribe linux-bcache" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Quoted 5 lines · On 12/15/2010 04:46 PM, [email protected] wrote:
First of all, I would like to thanks for answers.
One more question and some reply in text.
After making ssd disc a cache for some partition, there is possibility to detach cache. What should I do to "erase" (trim) whole disk for another benchmarking? I think that "dd if=/dev/zero of=/dev/sdx" wouldnt be right command.
Yeah, one nifty thing is you can disable caching on a mounted, in use filesystem (and reenable it provided you invalidate the cache somehow).
Hdparm can send a trim, but it can't do the whole disk at once, you've got to script it. I wrote one awhile back, this is what I use:
#!/bin/bash
DEV=$1
if [ ! -b "$DEV" ]; then
echo "Can't open $DEV"
exit 1
fi
SECTORS=hdparm -I "$DEV"|awk '/LBA48/ { print $5 }' OFFSET=0
echo "$DEV has $SECTORS sectors:"
while (($SECTORS)); do
i=$(( $SECTORS > 65535 ? 65535 : $SECTORS ))
echo "wiping $i sectors at offset $OFFSET"
hdparm --please-destroy-my-drive \
--trim-sector-ranges $OFFSET:$i $DEV
SECTORS=$(($SECTORS - $i))
OFFSET=$(($OFFSET + $i))done
--
To unsubscribe from this list: send the line "unsubscribe linux-bcache" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html