cp1
|
|
January 04, 2014, 05:28:59 PM |
|
After running for a couple of hours armory has failed to import my bitcoin wallet created with a bitcoin prior to 0.8.6 and later taken over by bitcoin 0.8.6. Now armory fails to import this wallet. Should I create a new wallet or try something else to import my wallet located in ~/.bitcoin?
I don't think armory is compatible with bitcoin-qt wallets. You should just create an armory wallet and send all your coins over to it.
|
|
|
|
arcke
Newbie
Offline
Activity: 46
Merit: 0
|
|
January 04, 2014, 05:30:01 PM |
|
OK
|
|
|
|
etotheipi (OP)
Legendary
Offline
Activity: 1428
Merit: 1093
Core Armory Developer
|
|
January 05, 2014, 06:00:56 PM |
|
Updates to 0.91-dev branch: (just pushed) - I have drastically reduced the tracking and handling of zero-conf transactions. This solves 80% of the speed issues using Armory with wallets with 10k+ addresses. I now run with 30k addresses and it runs fine, although it freezes for about 1-3 sec after every new block. Better than 10-15 sec, though (which was causing inter-thread timeouts). I will be working on the per-block scanning ops to reduce it even further (Armory code base should be able to handle 1mil+ addresses, I just haven't optimized the operations to do it)
- Updated "Send Bitcoins" dialog, with wallet selection directly in the dialog. Coin control still works, though the interface updates for that broke. Fixing that soon.
- Added paranoid-extension of address chains and logging of multipliers. All address chain extensions are now performed twice and answers are compared to make sure it's consistent. Additionally, the final multiplier/exponent applied is logged.
- There are still situations where an unclean shutdown will induce a rescan. I have further reduced those situations, though it's not perfect. I will see if the operations are efficient enough to make more reliable.
- armoryd.py updates -- a bunch of functions that broke in the update to 0.90 have been fixed. Most importantly, "listtransactions" and "getledger" now work. CircusPeanut is working on building a better test environment for it to improve the reliability.
- Updated leveldb block size to 32kb. This seems to improve scan speed on Linux/Mac, without breaking build speed on Windows. Be sure to --rebuild if you want the benefits of the new block size.
|
|
|
|
bitpop
Legendary
Offline
Activity: 2912
Merit: 1060
|
|
January 05, 2014, 06:58:57 PM |
|
Amazing. Does anyone know why my bitcoind stops answering after a few starts of armory?
|
|
|
|
cypherdoc
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
January 05, 2014, 07:42:07 PM |
|
i've not run into that. what is secure printing?
If you print on a modern printer with a hard drive, it will save a copy of everything you print that can later be stolen when the printer is sold. So secure print gives you a code to hand write on the page you print out -- both are needed to unlock the wallet. interesting. didn't know printers had a hard drive that could store a privkey. always thought this problem was isolated to the printing drum. what do you mean by "both are needed"? what's the use of printing anything out if you're going to handwrite out the chain code and seed anyways?
|
|
|
|
bitpop
Legendary
Offline
Activity: 2912
Merit: 1060
|
|
January 05, 2014, 07:50:46 PM |
|
i've not run into that. what is secure printing?
If you print on a modern printer with a hard drive, it will save a copy of everything you print that can later be stolen when the printer is sold. So secure print gives you a code to hand write on the page you print out -- both are needed to unlock the wallet. interesting. didn't know printers had a hard drive that could store a privkey. always thought this problem was isolated to the printing drum. what do you mean by "both are needed"? what's the use of printing anything out if you're going to handwrite out the chain code and seed anyways? You only write a small code
|
|
|
|
etotheipi (OP)
Legendary
Offline
Activity: 1428
Merit: 1093
Core Armory Developer
|
|
January 05, 2014, 07:54:49 PM |
|
what do you mean by "both are needed"? what's the use of printing anything out if you're going to handwrite out the chain code and seed anyways?
The SecurePrint code is substantially smaller than the rest of the data that was printed. I struggled with this a bit, because I needed the code to be long enough to be secure, but not so long that the user might as well just copy the rootkey & chaincode by hand anyway. So I struggled with this balance and settled for 8 bytes with super key-stretching. So that's 64 bits, plus approximately 20 bits of stretching (since the stretching will slow down guessing by about a factor of 1,000,000. Additionally, it requires 16 MB of RAM so it should be difficult for GPUs to parallelize it. It's not as strong as I'd prefer, but it's infeasible for any system to brute force right now. If you are ultra ultra-paranoid, you can just just copy everything by hand. Just make sure to use the backup tester before you store it.
|
|
|
|
cp1
|
|
January 05, 2014, 09:21:27 PM |
|
interesting. didn't know printers had a hard drive that could store a privkey. always thought this problem was isolated to the printing drum.
Just the large office-copier type ones. I remember a story where a police station sold their copier and the news station that bought it downloaded tons of personal criminal information off the hard drive. I'm sure it can be recovered from your local printer spool too.
|
|
|
|
HeliKopterBen
|
|
January 06, 2014, 02:57:23 AM |
|
interesting. didn't know printers had a hard drive that could store a privkey. always thought this problem was isolated to the printing drum.
Just the large office-copier type ones. I remember a story where a police station sold their copier and the news station that bought it downloaded tons of personal criminal information off the hard drive. I'm sure it can be recovered from your local printer spool too. Because of user demand, spooled data on newer models is often erased or scrammbled, but I wouldn't trust it with a private key. The best bet, especially if you are storing a lot of wealth, is to get a cheap dedicated USB inkjet printer and only connect it when you are printing paper wallets... or use the secure print feature.
|
Counterfeit: made in imitation of something else with intent to deceive: merriam-webster
|
|
|
cypherdoc
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
January 06, 2014, 04:28:54 AM |
|
any plans to get BIP 38?
|
|
|
|
arcke
Newbie
Offline
Activity: 46
Merit: 0
|
|
January 06, 2014, 05:38:03 PM |
|
I couldnt find this one on git: http://pastebin.com/p2eHL6qUJust happened to me after armory has been running for 2-3 days.
|
|
|
|
etotheipi (OP)
Legendary
Offline
Activity: 1428
Merit: 1093
Core Armory Developer
|
|
January 06, 2014, 09:59:45 PM |
|
Do you have a lot of addresses in Armory? This frequently occurs when Armory hits the 10-20 sec timeout during an operation... it happens most frequently when there is network burp, or you have 10k+ addresses and it's trying to update all of them after a new block comes in. If it's an address problem, we'll have a fix for it soon ... the 0.91-dev branch already has updates that make Armory operate much better with 50k+ addresses, though there's still more work to do.
|
|
|
|
RoadStress
Legendary
Offline
Activity: 1904
Merit: 1007
|
|
January 07, 2014, 12:10:55 AM |
|
Hey etotheipi, i tried running 2 instances of Armory (both with it's own wallet folder aka separate "--datadir" arguments, but same "--satoshidir" argument) and it doesn't work. How hard is it to implement multiple instances? If easy will you do it?
|
|
|
|
arcke
Newbie
Offline
Activity: 46
Merit: 0
|
|
January 07, 2014, 12:35:15 AM |
|
Do you have a lot of addresses in Armory? This frequently occurs when Armory hits the 10-20 sec timeout during an operation... it happens most frequently when there is network burp, or you have 10k+ addresses and it's trying to update all of them after a new block comes in. If it's an address problem, we'll have a fix for it soon ... the 0.91-dev branch already has updates that make Armory operate much better with 50k+ addresses, though there's still more work to do. I have 3 addresses right now. I don't think this is a problem. Might it be that it happened when I accidentally pulled my network cable out.
|
|
|
|
bitpop
Legendary
Offline
Activity: 2912
Merit: 1060
|
|
January 07, 2014, 02:03:30 AM |
|
Do you have a lot of addresses in Armory? This frequently occurs when Armory hits the 10-20 sec timeout during an operation... it happens most frequently when there is network burp, or you have 10k+ addresses and it's trying to update all of them after a new block comes in. If it's an address problem, we'll have a fix for it soon ... the 0.91-dev branch already has updates that make Armory operate much better with 50k+ addresses, though there's still more work to do. I have 3 addresses right now. I don't think this is a problem. Might it be that it happened when I accidentally pulled my network cable out. Sherlock Holmes
|
|
|
|
arcke
Newbie
Offline
Activity: 46
Merit: 0
|
|
January 07, 2014, 03:52:46 AM |
|
Do you have a lot of addresses in Armory? This frequently occurs when Armory hits the 10-20 sec timeout during an operation... it happens most frequently when there is network burp, or you have 10k+ addresses and it's trying to update all of them after a new block comes in. If it's an address problem, we'll have a fix for it soon ... the 0.91-dev branch already has updates that make Armory operate much better with 50k+ addresses, though there's still more work to do. I have 3 addresses right now. I don't think this is a problem. Might it be that it happened when I accidentally pulled my network cable out. Sherlock Holmes Watson
|
|
|
|
RAJSALLIN
|
|
January 08, 2014, 08:33:41 AM |
|
any plans to get BIP 38?
This. Think it's important that bip38 infrastructure is increased since it's one of the safest ways to store paper wallets.
|
|
|
|
Ente
Legendary
Offline
Activity: 2126
Merit: 1001
|
|
January 08, 2014, 10:13:03 AM |
|
any plans to get BIP 38?
This. Think it's important that bip38 infrastructure is increased since it's one of the safest ways to store paper wallets. This sounds like the same thing: what do you mean by "both are needed"? what's the use of printing anything out if you're going to handwrite out the chain code and seed anyways?
The SecurePrint code is substantially smaller than the rest of the data that was printed. I struggled with this a bit, because I needed the code to be long enough to be secure, but not so long that the user might as well just copy the rootkey & chaincode by hand anyway. So I struggled with this balance and settled for 8 bytes with super key-stretching. So that's 64 bits, plus approximately 20 bits of stretching (since the stretching will slow down guessing by about a factor of 1,000,000. Additionally, it requires 16 MB of RAM so it should be difficult for GPUs to parallelize it. It's not as strong as I'd prefer, but it's infeasible for any system to brute force right now. If you are ultra ultra-paranoid, you can just just copy everything by hand. Just make sure to use the backup tester before you store it. It isn't, by chance, BIP38 or compatible? Ente
|
|
|
|
bitpop
Legendary
Offline
Activity: 2912
Merit: 1060
|
|
January 08, 2014, 11:43:05 PM |
|
|
|
|
|
etotheipi (OP)
Legendary
Offline
Activity: 1428
Merit: 1093
Core Armory Developer
|
|
January 09, 2014, 12:54:43 AM |
|
Armory already achieves the same security properties of BIP 38, just with a different algorithm. If BIP 38 becomes a real standard, I'll be happy to implement, though I don't see any deficiencies in my current algorithm except for interoperability with other apps. If everyone else is committed to BIP 38, I can be too.
|
|
|
|
|