Error Connecting to MySQL DB 4.8.1.1
El Cas
eaglexboy at gmail.com
Mon Feb 26 03:37:31 UTC 2018
I was able to use KMM 4.8.1.1 for about 4hr before my box became
unresponsive. I don't know if it was due to a memory leak or something
else. Nothing showed in the output and I was not really paying attention to
what top was reporting regarding memory usage, I was focusing on cpu usage.
During the whole time KMM peaked at 30% when it was processing the update
request I made.
I entered some transaction and noticed a discrepancy in speeds. When I
entered a transaction from the scheduled section it took about 3 sec but
when I entered a transaction via the ledger section it took about 6
sections per entry. When I imported a OFX file I downloaded from my bank it
took a while to process the small file (did not time the process) but when
I did an update from KMM against my bank account it took almost 2 hours to
process the 124 entries it downloaded. When I use to do this with 4.6 it
use to take 10 to 15 mins.
I uninstall KMM and re-install it using the package manager (which is
4.8.0-2) and see if I notice any difference in speed or if my system become
unresponsive with this version but turns out I couldn't use it since it has
the bug where the table kmmPayeesPayeeIdentifier has a field called *order*,
but order is a MySQL keyword. I even tried renaming the field just to see
if it would work but it did not. So I just reinstalled 4.8.1.1 and
continued using it. This time I will send top's output to a file so I can
see the memory and CPU usage if I my box becomes unresponsive again.
I also tried to build KMM 5.0 but but it doesn't seem to be able to find
libalkimia5 (I have both the normal and dev installed). I did not have to
set it anything for it when I build 4.8.1.1 from source. Do I just locate
the lib file and set *LibAlkimia5_DIR* to it when I run cmake or do I need
to do something else? (I'm still a noob when it comes to building from
source)
*CMake Error at CMakeLists.txt:58 (find_package):By not providing
"FindLibAlkimia5.cmake" in CMAKE_MODULE_PATH this project has asked CMake
to find a package configuration file provided by "LibAlkimia5", but CMake
did not find one. Could not find a package configuration file provided by
"LibAlkimia5" (requested version 7.0.0) with any of the following
names: LibAlkimia5Config.cmake libalkimia5-config.cmake Add the
installation prefix of "LibAlkimia5" to CMAKE_PREFIX_PATH or set
"LibAlkimia5_DIR" to a directory containing one of the above files. If
"LibAlkimia5" provides a separate development package or SDK, be sure it
has been installed.*
On Sun, Feb 25, 2018 at 4:37 PM, El Cas <eaglexboy at gmail.com> wrote:
> It appears it finished processing my entries. I was able to start up the
> application and it logged in and loaded my database.
>
> First off, thanks for adding the Tags feature as adding the tags in memo
> was not the best way to search and group transactions. The one thing I have
> noticed is that it's take a bit longer to switch between sections and
> accounts in the ledger compared to 4.6.6 (from ~2 sec to ~6sec). I have the
> filter set to only show the past two months. To see if it would make a
> difference I changed the filter to the start of month (so it only show the
> past 25 days) but the response time remained about the same. The only thing
> I saw in the terminal was *KMyMoneyPlugin::PluginLoader::
> createInstanceFromService returned 0 for SEPA orders with error Could not
> find plugin 'SEPA orders' for application 'kmymoney' *when I updated the
> filter. I don't know if this delay is due to the way it communicates with
> DB now vs 4.6.6 or if something else is going on. I made sure to not have
> anything else running that would make it appear that KMM was running slow.
> At its peak KMM used 8% of my CPU and 137Mb of RAM.
>
> System specs
> Processor: AMD Fx-4350 Quad Core
> Memory: 32Gb Ram, 64 Gb Swap
> OS: Kubuntu 17.10 (Kernel 4.13.0-36)/64-Bit
>
> Is there something I can do to speed up or check why it's taking longer
> then normal?
>
>
> On Sun, Feb 25, 2018 at 3:39 PM, El Cas <eaglexboy at gmail.com> wrote:
>
>> So after running for 11 hours my box became unresponsive. I could move
>> the mouse and the cursor reflected the move but that was all I could do. I
>> tried to SSH and I got a response to the connect attempt but never got
>> prompted for password (which would make sense seeing how I could only use
>> the mouse on it).
>>
>> I have 32,248 rows in *kmmsplits* which is a good amount but I don't
>> know how exactly the code cycles through them during 2 < 4. If it loads
>> them in memory then this would explain why my box appeared to be
>> unresponsive. I tried to build it with debug enabled but I can't seem to
>> build i. I keep getting the following two errors.
>>
>> /kmm/kmymoney/libkdchart/src/KDChart/Cartesian/KDChartCartesianAxis.cpp:
>> In member function ‘void KDChart::TickIterator::compute
>> MajorTickLabel(int)’:
>> /mnt/WorkingFolder/temp/kmm/kmymoney/libkdchart/src/KDChart/
>> Cartesian/KDChartCartesianAxis.cpp:276:46: *error*: conversion from
>> ‘QMap<double, QString>::iterator’ to non-scalar type ‘QMap<double,
>> QString>::ConstIterator {aka QMap<double, QString>::const_iterator}’
>> requested
>>
>> *m_dataHeaderLabels.lowerBound( slightlyLessThan( m_position )
>> );
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~*
>> /mnt/WorkingFolder/temp/kmm/kmymoney/libkdchart/src/KDChart/
>> Cartesian/KDChartCartesianAxis.cpp: In member function ‘void
>> KDChart::TickIterator::operator++()’:
>> /mnt/WorkingFolder/temp/kmm/kmymoney/libkdchart/src/KDChart/
>> Cartesian/KDChartCartesianAxis.cpp:306:76: *error*: conversion from
>> ‘QMap<double, QString>::iterator’ to non-scalar type ‘QMap<double,
>> QString>::ConstIterator {aka QMap<double, QString>::const_iterator}’
>> requested
>>
>> *QMap< qreal, QString >::ConstIterator it = m_annotations.upperBound(
>> m_position );
>> ~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~*
>>
>> I ran *cmake -D CMAKE_BUILD_TYPE=Debugfull ..* and *make* as the
>> README.cmake said to do. Am I missing something?
>>
>> Is there another way to monitor what is going on during the looping of
>> records in 2 < 4?
>>
>> While I await for a response I will try to run it again and direct all
>> output to a file so that if it hangs again I can at least see what it spit
>> out on the terminal.
>>
>>
>> On Sun, Feb 25, 2018 at 9:53 AM, Thomas Baumgart <thb at net-bembel.de>
>> wrote:
>>
>>>
>>> On Sonntag, 25. Februar 2018 04:51:18 CET El Cas wrote:
>>>
>>> > On Sun, Feb 25, 2018 at 2:31 AM, Thomas Baumgart <thb at net-bembel.de>
>>> wrote:
>>> > > Having the DB on a separate server should not make a difference. All
>>> > > communication (incl. structure updates etc) between KMyMoney and the
>>> SQL
>>> > > DB is
>>> > > via a communication link.
>>> > >
>>> > > I am not sure, if bringing in 5.0 at this point is a good
>>> alternative. You
>>> > > might have a trouble to go back to 4.8 unless you want to stick with
>>> 5.0.
>>> > > Since you're building from source also, use the 5.0 branch, not
>>> master. It
>>> > > already contains a couple of bug-fixes over the released 5.0 version.
>>> >
>>> > The reason I was using Kubuntu 18.04/KMM 5.0 was because 18.04 is
>>> > scheduled to be released in April and if this fixed my problem I would
>>> just
>>> > going to completely upgrade my system to 18.04/5.0 ahead of scheduled.
>>> >
>>> > So while using Kubuntu 18.04/KMM 5.0 I was getting a database user
>>> error so
>>> > I just temporarily granted the username I am using for KMM all
>>> privileges.
>>> > I created an empty test account and proceeded to save as DB. I put in
>>> the
>>> > required information and when I clicked on Ok. Got the error below.
>>> >
>>> > *Error in function int MyMoneyStorageSql::open(const QUrl&, int, bool)
>>> :
>>> > opening new database*
>>> >
>>> > *Driver = QMYSQL, Host = X.X.X.X, User = kmymoney, Database =
>>> /newkmymoney*
>>> >
>>> > *Driver Error: QMYSQL: Unable to connect*
>>> >
>>> > *Database Error No 1049: Unknown database '/newkmymoney'*
>>> >
>>> > *Text: Unknown database '/newkmymoney' QMYSQL: Unable to connect*
>>> >
>>> > *Error type 1*
>>> >
>>> > *Executed: *
>>> >
>>> > *Query error No -1: *
>>> >
>>> > *Error type 0*
>>> >
>>> > I checked the DB server and the Database was created but it was empty.
>>> >
>>> > So I went back to Kubuntu 17.10/KMM 4.8.1.1 (from source) and repeated
>>> the
>>> > same process. This time the test database was created and the tables
>>> were
>>> > created. I decided to attempt to open my original database and got the
>>> same
>>> > error I reported in my first email.
>>> >
>>> > As for difference between 4.8.1.1 vs 4.6.6 I found the following
>>> >
>>> > 4.6 has 24 tables vs 23 tables on 4.8: *kmmtmpfileinfo*
>>>
>>> That's not a problem. This seems to be some leftover from a previous run
>>> (probably the creation of the new columns in kmmfileinfo). You should be
>>> able
>>> to drop that table.
>>>
>>> > And the following tables have extra columns
>>> > Version tablename field name
>>> > type Null? Default Value
>>> > 4.8 *kmmfileinfo* hiOnlineJobId bigint(20) unsigned
>>> > YES NULL
>>> > 4.8 *kmmfileinfo* hiPayeeIdentifierId bigint(20) unsigned
>>> > YES NULL
>>> > 4.8 *kmmschedules* lastDayInMonth char(1)
>>> > NO NULL
>>> > 4.8 *kmmsplits* costCenterId varchar(32)
>>> > YES NULL
>>>
>>> Those are the new columns that were added in 4.8. So that looks about
>>> right
>>> and 4.8 cannot start w/o them.
>>>
>>> > I made a back up of my current DB and added the missing fields to the
>>> > corresponding tables and I was able to open the Database. It proceeded
>>> load
>>> > and when it reached 97% it began to do *testing fileFixVersion*.
>>> During the
>>> > 0 < 4 run it performed several *Fixed interest action in XXXX*
>>> actions. It
>>> > then ran *testing fileFixVersion* *1 < 4 and 2 <4 *has been running
>>> for a
>>> > long time (almost 2hr as of this email). I don't know how long the
>>> testing
>>> > normally takes but I will let it continue running and see how long it
>>> takes
>>> > to finish.
>>> >
>>> > I am going to check on it when I wake up. I will report any progress,
>>> or
>>> > lack there off.
>>>
>>> filefix 2 < 4 is doing the following according to the source code:
>>>
>>> // scan the transactions and modify transactions with two splits
>>> // which reference an account and a category to have the memo text
>>> // of the account.
>>>
>>> filefix 3 < 4 is doing the following:
>>>
>>> // make sure each storage object contains a (unique) id
>>>
>>> That is just a single operation and should not take long.
>>>
>>>
>>> I can't tell you anything about the speed. Usually I use KMyMoney with a
>>> GPG
>>> encrypted XML file and not a database backend.
>>>
>>>
>>> --
>>>
>>> Regards
>>>
>>> Thomas Baumgart
>>>
>>> https://www.telegram.org/ Telegram, the better WhatsApp
>>> -------------------------------------------------------------
>>> You cannot strengthen the weak by weakening the strong.
>>> -------------------------------------------------------------
>>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kmymoney-devel/attachments/20180225/97895d26/attachment-0001.html>
More information about the KMyMoney-devel
mailing list