Armory 0.96.3 released

[Holliday]

Ah, neat!
I upgraded both ArmoryDB and ArmoryQt to 0.96.4, still the exact same messages and scanning never finishes..

Ente

Did you happen to delete the armory/databases folder after updating and try the new version from scratch?

[goatpig]

Ah, neat!
I upgraded both ArmoryDB and ArmoryQt to 0.96.4, still the exact same messages and scanning never finishes..

Ente

I need to see the full Qt log.

[goatpig]

i also tried building from the 0.96.4 source but run into this error upon "make":

Code:

git submodule init
git submodule update
sh autogen.sh
./configure
make clean
make -j8
sudo make install

[Ente]

i also tried building from the 0.96.4 source but run into this error upon "make":

Code:

user@user-VirtualBox:~/Downloads/BitcoinArmory-0.96.4$ make
Making all in cppForSwig
make[1]: Entering directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig'
Making all in lmdb
make[2]: Entering directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig/lmdb'
make[2]: Nothing to be done for 'all'.
make[2]: Leaving directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig/lmdb'
Making all in fcgi
make[2]: Entering directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig/fcgi'
make[2]: *** No rule to make target 'all'.  Stop.
make[2]: Leaving directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig/fcgi'
Makefile:1538: recipe for target 'all-recursive' failed
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory '/home/user/Downloads/BitcoinArmory-0.96.4/cppForSwig'
Makefile:413: recipe for target 'all-recursive' failed
make: *** [all-recursive] Error 1

before

sh autoconfigure.sh

do

git submodule init
git submodule update

edit: oops, didn't notice the new thread page.

Cheers,

Ente

[Ente]

Ah, neat!
I upgraded both ArmoryDB and ArmoryQt to 0.96.4, still the exact same messages and scanning never finishes..

Ente

I need to see the full Qt log.

Is ArmoryQT --debug enough? Nothing more to see, afaik:

(INFO) ArmoryQt.py:566 - Usermode: Expert
(INFO) ArmoryQt.py:1698 - Changing usermode:
(INFO) ArmoryQt.py:1699 -    From: Expert
(INFO) ArmoryQt.py:1707 -      To: Expert
(INFO) ArmoryQt.py:1836 - startBitcoindIfNecessary
(WARNING) ArmoryQt.py:1826 - DB is already running
(INFO) ArmoryQt.py:1924 - Setting netmode: 1
-ERROR - : (BinaryData.h:563) odd hexit count
-ERROR - : (SwigClient.cpp:61) odd hexit count
(INFO) ArmoryQt.py:1924 - Setting netmode: 0
(INFO) ArmoryQt.py:1906 - loadBlockchainIfNecessary
(INFO) ArmoryQt.py:4673 - Dashboard switched to "Scanning" mode
(INFO) ArmoryQt.py:4673 - Dashboard switched to "Scanning" mode

Cheers,

Ente

[goatpig]

I gather you're getting to load the client before it gets to that stage, right? Have you tried deleting the .lmdb mirror files? Also, consider running the wallet through the recovery tool in full mode.

[alomar]

i also tried building from the 0.96.4 source but run into this error upon "make":

Code:

git submodule init
git submodule update
sh autogen.sh
./configure
make clean
make -j8
sudo make install

since i don't use git, how can i accomplish the same thing?  i merely downloaded the source code tarball zip directly from your github, unpacked, and ran the standard commands.

[goatpig]

since i don't use git, how can i accomplish the same thing?  i merely downloaded the source code tarball zip directly from your github, unpacked, and ran the standard commands.

Which source? The link called "Source Code"? That's an automated tarball put together by GitHub, which somehow skips submodules (don't ask why, I wish I knew). You want to use the signed sources. There is only a signed tag for 0.96.4 atm, so you'd have to use git to procure that code. There's a signed source for RC3 however:

https://github.com/goatpig/BitcoinArmory/releases/download/v0.96.3.992/armory_0.96.3.992_src.tar.gz

[Ente]

I gather you're getting to load the client before it gets to that stage, right? Have you tried deleting the .lmdb mirror files? Also, consider running the wallet through the recovery tool in full mode.

I removed the two new wallets (one imported watch-only, one freshly created with one imported address), and removed all .lmdb and .lmdb-lock files. And I checked all wallets on "full recovery", no errors found.
Still same behavior.
Like, now, I only have my ages old walletsagain, and still have the "odd hexit count" errors in Qt.

edit: restarting ArmoryDB, and everything works with the old wallets only. I'll check a few combinations of wallets, files, scans, and will report back.

Is this a wallet problem? And/or a Qt problem? Or a DB problem? DB seems to start scanning from the beginning every time too.

Ente

[goatpig]

I'm thinking it's a Qt problem.

[Ente]

So, with the newly generated wallet including one imported address, it scans history for an hour, then DB hangs, at -30secs, with

(BlockchainScanner.cpp:852) scanned from block #514744 to #515526

as the last line in dbLog.txt. I can't get a Qt connection any more after restarting Qt, no matter what wallets it has loaded.
After restarting DB, Qt connects fine and loads wallets and tx instantly.

With the watch-only wallet newly added, exactly the same as above.

To triplecheck, after removing those wallets again, I created a fresh wallet while not connected to DB. Restarting DB and then Qt, it starts scanning history. Everything again the same, seconds before finishing it hangs DB and Qt. This time I got ever-repeating errors in Qt:

(python:7717): Gtk-CRITICAL **: IA__gtk_progress_configure: assertion 'value >= min && value <= max' failed
-ERROR - : (BinaryData.h:563) odd hexit count
-ERROR - }: (BinaryData.h:563) odd hexit count
-ERROR - : (BinaryData.h:563) odd hexit count
-ERROR - }: (BinaryData.h:563) odd hexit count

Well, I have no clue how to go on from here?

Ente

[goatpig]

To triplecheck, after removing those wallets again, I created a fresh wallet while not connected to DB. Restarting DB and then Qt, it starts scanning history. Everything again the same, seconds before finishing it hangs DB and Qt. This time I got ever-repeating errors in Qt:

You mean the DB hangs even without any imports?

[Ente]

To triplecheck, after removing those wallets again, I created a fresh wallet while not connected to DB. Restarting DB and then Qt, it starts scanning history. Everything again the same, seconds before finishing it hangs DB and Qt. This time I got ever-repeating errors in Qt:

You mean the DB hangs even without any imports?

These symptoms happen with creating a new plain armory-wallet via GUI, yes. Nothing imported, neither the wallet itself nor any single address.
Of course I kept Armory offline when creating the wallet, so it would start a full rescan when being online, reaching DB again.

Ente

[goatpig]

Start the DB on its own (no GUIs), how far does it get? Try and post the last 10 lines of the dbLog.

[Ente]

ArmoryDB is running by itself on its own headless host.
Restarting the ArmoryDB service runs just fine, the last dbLog.txt entries are:

-WARN  - 16:13:54: (BDM_supportClasses.cpp:1944) running 35 zc parser threads
-INFO  - 16:21:05: (BlockchainScanner.cpp:857) scanned block #518330
-INFO  - 16:21:26: (BlockchainScanner.cpp:857) scanned block #518331

Starting ArmoryQt on a remote machine:
(again, dbLog.txt)

-INFO  - 16:28:37: (BDM_Server.cpp:1121) registered bdv: a*****0
-INFO  - 16:28:38: (BDM_supportClasses.cpp:401) Starting address registration process
-INFO  - 16:28:43: (BlockchainScanner.cpp:852) scanned from block #0 to #142691
[..]

and then, finally:

-INFO  - 17:11:48: (BlockchainScanner.cpp:852) scanned from block #514744 to #515526

As the last line, from then on ArmoryDB hangs.

ArmoryQt says:

(INFO) ArmoryQt.py:4673 - Dashboard switched to "Scanning" mode
-ERROR - : (BinaryData.h:563) odd hexit count

(scanning comes a few hundred times, then the hexit error several times a second)

Ente

[goatpig]

The DB runs fine. The issue is the client. Do you have comments in that wallet, written in a non ASCII locale?

[Ente]

To remove factors, I tried again:

- removed all wallets, ArmoryQt starts fine in online mode
- created a new wallet, it's online immediately
- shut down Qt and DB
- started Qt, created a new wallet with password, nothing else changed or touched (in offline mode)
- shutdown Qt
- started DB, then Qt
- Qt starts scanning

(INFO) ArmoryQt.py:4673 - Dashboard switched to "Scanning" mode
(python:14412): Gtk-CRITICAL **: IA__gtk_progress_configure: assertion 'value >= min && value <= max' failed

"Scanning" comes around every 5 secs, that Gtk warning came only once. Nothing else for an hour scanning.

At "35 sec left", the last DBlog.txt line is

-INFO  - 12:29:15: (BlockchainScanner.cpp:852) scanned from block #514744 to #515526

(which is not the current blockheight)

and Qt says

-ERROR - : (BinaryData.h:563) odd hexit count

over and over.

After restarting Qt, I can't connect to DB any more, I have to restart the service.

tl;dr: a single fresh plain wallet, nothing changed, nor commented, and Qt hangs.

Oh, maybe it has something to do with a recent upgrade to Debian 9? I probably didn't import/rescan since then.
But I did install/compile v0.96.4, which threw no errors.

Ente

[goatpig]

To remove factors, I tried again:

- removed all wallets, ArmoryQt starts fine in online mode
- created a new wallet, it's online immediately
- shut down Qt and DB
- started Qt, created a new wallet with password, nothing else changed or touched (in offline mode)
- shutdown Qt
- started DB, then Qt
- Qt starts scanning

(INFO) ArmoryQt.py:4673 - Dashboard switched to "Scanning" mode
(python:14412): Gtk-CRITICAL **: IA__gtk_progress_configure: assertion 'value >= min && value <= max' failed

"Scanning" comes around every 5 secs, that Gtk warning came only once. Nothing else for an hour scanning.

At "35 sec left", the last DBlog.txt line is

-INFO  - 12:29:15: (BlockchainScanner.cpp:852) scanned from block #514744 to #515526

(which is not the current blockheight)

and Qt says

-ERROR - : (BinaryData.h:563) odd hexit count

over and over.

After restarting Qt, I can't connect to DB any more, I have to restart the service.

tl;dr: a single fresh plain wallet, nothing changed, nor commented, and Qt hangs.

Oh, maybe it has something to do with a recent upgrade to Debian 9? I probably didn't import/rescan since then.
But I did install/compile v0.96.4, which threw no errors.

Ente

Send me one of those fresh, broken wallets.

[Ente]

Send me one of those fresh, broken wallets.

E-mail sent.

Ente

[Ente]

I started over with a freshly installed Debian 9, and then with again with a completely empty .armory folder too. v0.96.4, local ArmoryQT connecting to remote ArmoryDB.
Again, the same symptoms, it stops scanning at "one minute left" with those hexit errors.

I did everything exactly like always, with dependencies and commands from old written down notes.
So it can only be something broken at my ArmoryDB, or a general imcompatibility with the Qubes OS Debian VM templates, or a networking problem in between?

Ente