Author

Topic: [XMR] Monero - A secure, private, untraceable cryptocurrency - page 935. (Read 4670673 times)

full member
Activity: 186
Merit: 100
Monero
From the other side of town:

https://bitseed.org/product/pre-order-bitseed-v2/

Quote
Also new for Bitseed v2 is support for altcoins – Bitseed Altcoin Edition is now available for pre-order!

We will be handling altcoins on a per-coin basis, and will only be accepting orders for single altcoin devices after the first 100 unit sales of devices supporting a particular altcoin e.g. if we receive an order for 100 units to support Litecoin, we will then add the option to order single units that support Litecoin. We encourage any interested altcoin communities to put together a group order of 100 units and then send us a message to receive an invoice for your bulk order. Be sure to note which software you want us to add support for and send a spreadsheet or some other kind of organized document with the individual order information (shipping details and unit amounts).

So it would take 12k to get the 100 units.



Are these just for running a full node? Doesn't seem to be quite worth it to me compared to a Rasp pi unless there is something else they can offer.
legendary
Activity: 3836
Merit: 4969
Doomed to see the future and unable to prevent it
I'm not familiar with IOS file system but does it support directory compression?
legendary
Activity: 2968
Merit: 1198
Yes it's just not automatic the way it is in Linux. There are some other issues with it, too, but nothing show stopping. Also, sparse files aren't supported at all on the Mac. Anyway, we'll get it worked out.

The Mac thing is a little deeper/weirder than that. Here's what I mean:


[ ric->lmdb ]$ ruby
File.open('bigassfile_DELETE_ME', 'w') do |f|
    f.seek(1000000000)
    f.puts('l')
end
[ ric->lmdb ]$ ls -lh
total 14107248
-rw-r--r--  1 ric  staff   954M May  7 16:24 bigassfile_DELETE_ME
-rw-r--r--  1 ric  staff   5.8G May  7 16:23 data.mdb
-rw-r--r--  1 ric  staff   8.0K May  7 09:27 lock.mdb


In other words, the leading nulls in the 1gb file count as actual data (whereas ext2 and above would let you use that space and would only allocate blocks as needed). However, as you can see, the LMDB file doesn't have the full 16gb allocated, which (I would imagine) is the OS X VFS handling the sparse file abstraction.

There are two different things going on. One is sparse files, which aren't supported on OS X (unless you are using a non-standard file system). The other is whether lmdb extends the database file to the maximum size or not. It DOES do this on Windows and DOES NOT do this on other operating systems. The reason is not clear (and the documentation is terrible).

(If you do the above test on Linux -- which does support sparse files for typical file systems -- you will see the same result except for the first line (total XXX) which shows actual on-disk usage.)

legendary
Activity: 1708
Merit: 1049
Yesterday, while searching some other stuff, I bumped into zram for linux: http://en.wikipedia.org/wiki/Zram

This can be handy for Monero if some older linux boxes have limited RAM. In general purpose use I'm seeing compression ratios of 2-3x, meaning that a single gb of ram is turned into 2-3gb. It has a tradeoff with cpu power (as it requires cpu to compress/decompress the data on the fly) but the cpu will be much faster than swapping data to the hard disk. It's a no brainer.
legendary
Activity: 2282
Merit: 1050
Monero Core Team

The bitcoin network has become sentient and is attacking it's enemies.

No but seriously it works fine for me.

Actually the link works now, but for a few moments well ...
legendary
Activity: 2282
Merit: 1050
Monero Core Team
donator
Activity: 1274
Merit: 1060
GetMonero.org / MyMonero.com
Yes it's just not automatic the way it is in Linux. There are some other issues with it, too, but nothing show stopping. Also, sparse files aren't supported at all on the Mac. Anyway, we'll get it worked out.

The Mac thing is a little deeper/weirder than that. Here's what I mean:


[ ric->lmdb ]$ ruby
File.open('bigassfile_DELETE_ME', 'w') do |f|
    f.seek(1000000000)
    f.puts('l')
end
[ ric->lmdb ]$ ls -lh
total 14107248
-rw-r--r--  1 ric  staff   954M May  7 16:24 bigassfile_DELETE_ME
-rw-r--r--  1 ric  staff   5.8G May  7 16:23 data.mdb
-rw-r--r--  1 ric  staff   8.0K May  7 09:27 lock.mdb


In other words, the leading nulls in the 1gb file count as actual data (whereas ext2 and above would let you use that space and would only allocate blocks as needed). However, as you can see, the LMDB file doesn't have the full 16gb allocated, which (I would imagine) is the OS X VFS handling the sparse file abstraction.
legendary
Activity: 1260
Merit: 1008
From the other side of town:

https://bitseed.org/product/pre-order-bitseed-v2/

Quote
Also new for Bitseed v2 is support for altcoins – Bitseed Altcoin Edition is now available for pre-order!

We will be handling altcoins on a per-coin basis, and will only be accepting orders for single altcoin devices after the first 100 unit sales of devices supporting a particular altcoin e.g. if we receive an order for 100 units to support Litecoin, we will then add the option to order single units that support Litecoin. We encourage any interested altcoin communities to put together a group order of 100 units and then send us a message to receive an invoice for your bulk order. Be sure to note which software you want us to add support for and send a spreadsheet or some other kind of organized document with the individual order information (shipping details and unit amounts).

So it would take 12k to get the 100 units.

legendary
Activity: 2968
Merit: 1198
Problem is sparse either doesn't work on Windows (or Mac, I think), or doesn't work well, at least in some configurations. Nevertheless this will be addressed once the ability to expand the database is added. Then the starting size can be backed off., at least for non-sparse deployments.

Regarding sparse files, is this an LMDB issue or a more general file system issue?

There does seem to be sparse file support in windows, but it is dependent on the file system:

https://msdn.microsoft.com/en-us/library/windows/desktop/aa365564(v=vs.85).aspx
https://msdn.microsoft.com/en-us/library/windows/desktop/aa365566(v=vs.85).aspx
https://technet.microsoft.com/en-gb/library/cc788025.aspx
http://ntfs.com/ntfs-sparse.htm

Yes it's just not automatic the way it is in Linux. There are some other issues with it, too, but nothing show stopping. Also, sparse files aren't supported at all on the Mac. Anyway, we'll get it worked out.
sr. member
Activity: 280
Merit: 250
Problem is sparse either doesn't work on Windows (or Mac, I think), or doesn't work well, at least in some configurations. Nevertheless this will be addressed once the ability to expand the database is added. Then the starting size can be backed off., at least for non-sparse deployments.

Regarding sparse files, is this an LMDB issue or a more general file system issue?

There does seem to be sparse file support in windows, but it is dependent on the file system:

https://msdn.microsoft.com/en-us/library/windows/desktop/aa365564(v=vs.85).aspx
https://msdn.microsoft.com/en-us/library/windows/desktop/aa365566(v=vs.85).aspx
https://technet.microsoft.com/en-gb/library/cc788025.aspx
http://ntfs.com/ntfs-sparse.htm

HTH, Q
hero member
Activity: 658
Merit: 503
Monero Core Team
Updated my letter to an investor with #6.
Quote
In contrast, Bitcoin technology is getting easier and easier to track - #1, #2, #3, #4, #5, #6.

I might consider starting Monero meetup in Paris. Might.
legendary
Activity: 2968
Merit: 1198
I have berkeley db running on raspberry pi 2 but the sync time is incredibly slow.

Nevertheless, it runs and someday it will be a functioning member of node society.

Gotta let these youngsters grow!

Development is actively being done to speed up bdb specifically and all dbs.

legendary
Activity: 1750
Merit: 1036
Facts are more efficient than fud
LMDB version not official.
Blockchain conversion tools broken.
No official GUI.
No TAILS binaries available
Freebazaar nowhere near complete, last git submission 28 days ago.
DwarfPool scam ongoing, efficiency below 80%

I can't wait to see fluffyponny in person later this month, got a full bag of rotten eggs and tomatoes just for him..





awesome, i´ll also be there, hope u can run faster than i can throw a beer bottle.

I'll buy you a keg. Here's Atrides post yesterday from GLIP freebazaar group:


"Atrides FreeBazaar (dwarfpool)
I have merged OpenBazaar Beta4 and fixed conflicts"

Not sure on the entire status, but you see submissions from hours to a few days to a few weeks ago--maybe not a good idea to look at the first number you see and claim it's the last submission....  Roll Eyes

https://github.com/freebazaar/FreeBazaar
legendary
Activity: 1610
Merit: 1004
I have berkeley db running on raspberry pi 2 but the sync time is incredibly slow.

Nevertheless, it runs and someday it will be a functioning member of node society.

Gotta let these youngsters grow!

hero member
Activity: 826
Merit: 500
I see that one of the development goals is the implementation of BerkleyDB as a fallback. Will this be as time intensive as the work on LMDB has been?

It´s already implemented and works on Raspberry PI etc.

There are currently 3 database types with the DB interface:
- LMDB (normal pcs)
- Berkley DB (embedded devices like RPi)
- In Memory dB (not the old one, it also uses the DB interface, for whatever services who need the whole db in ram for fast searching etc.)
Good to hear. Thanks
hero member
Activity: 532
Merit: 500
I see that one of the development goals is the implementation of BerkleyDB as a fallback. Will this be as time intensive as the work on LMDB has been?

It´s already implemented and works on Raspberry PI etc.

There are currently 3 database types with the DB interface:
- LMDB (normal pcs)
- Berkley DB (embedded devices like RPi)
- In Memory dB (not the old one, it also uses the DB interface, for whatever services who need the whole db in ram for fast searching etc.)
hero member
Activity: 826
Merit: 500
I see that one of the development goals is the implementation of BerkleyDB as a fallback. Will this be as time intensive as the work on LMDB has been?
legendary
Activity: 2968
Merit: 1198
LMDB version not official.
Blockchain conversion tools broken.
No official GUI.
No TAILS binaries available
Freebazaar nowhere near complete, last git submission 28 days ago.
DwarfPool scam ongoing, efficiency below 80%

I can't wait to see fluffyponny in person later this month, got a full bag of rotten eggs and tomatoes just for him..

Agree, fluffypony moves like a true pony. Load him by heavy bag of dollars. Let pony to work Smiley

I see, many users report successful blockchain conversion to LMDB with batch size = 1000. The little issue is 16GB sparse database file. 'Sparceness' is well known feature to programmers, it's normal. Actual database file size is much less, say 3-5 Gb. Note memory fingerprint instead - users do report 30-60Mb.

IT IS THE WIN.

Problem is sparse either doesn't work on Windows (or Mac, I think), or doesn't work well, at least in some configurations. Nevertheless this will be addressed once the ability to expand the database is added. Then the starting size can be backed off., at least for non-sparse deployments.
sr. member
Activity: 350
Merit: 250
Belgium, France, Germany, Holland... Monero is on a tour

https://twitter.com/monerocurrency
Jump to: