Stability seems good to me. I sometimes get MS SQL and MySQL database
corruptions either from power outages or OOM. I think you've done an
insanely awesome job putting H2 together Thomas. Other more "Mainstream"
products seem to get more hype and attention, but I don't think that
makes them
Hi,
Can you try a unique constraint instead?
See here
http://osdir.com/ml/h2-database/2012-03/msg00141.html
Ryan
On 3/05/2016 10:36 PM, sd23gl via H2 Database wrote:
Hello,
I do save usernames in a VARCHAR_IGNORECASE column. However the user
can stay anonymous until a username is picked.
Hi,
You could make it work with that, but it would depend massively on how
you architect your system. You won't just be able to put the records in
a database and then query it with any kind of scalability and
performance expectations.
Ryan
On 20/04/2016 10:12 PM, Valentin Popov wrote:
You aren't executing the query, only creating it?
|session.createSQLQuery("BACKUP TO
'"+file.getCanonicalPath()+"'").executeUpdate();|
On 4/04/2016 4:55 AM, abdou amer wrote:
|I want to perform backup and restore using hibernate ORM, but i seems
that the below code not do any thing.|
I've been using MVCC with PageStore for years. LOBS don't work very well
(but I didn't want them in the db anyway, so we don't use LOBS). Haven't
had any issues with anything else though.
Doesn't mean bugs don't exist though.
You only get the issue with MVCC on though hey?
On 17/02/2016 1:18
That would be the job of the uninstaller. eg. If they are all in a
single directory then delete that directory.
So I'd look up the documentation for the installer / uninstaller.
On 19/01/2016 11:15 PM, AKHIL KOTTAKKAL AKHIL wrote:
how can i delete all .db files when i uniinstall my java
That's so good I just wanted to see it by itself and repeat it to everyone!
:D
On 27/10/2015 4:28 PM, Lukas Eder wrote:
As far as I'm concerned, nothing beats a good glass of red wine whilst
reading the SQL standard documents in front of the fireplace. Somehow,
few people are with me on that
You need to create the Kunde table before the |RESERVIERUNG table.
Otherwise it can't create the foreign key (because the table doesn't exist).
|
On 15/10/2015 6:04 PM, Steve McLeod wrote:
Two things I can suggest:
1) You should ask this question on stackoverflow.com, not in this
mailing
I'd be inclined to have a connection pool per session. Then create and
destroy the connection pool. That way it will manage connections however
you want to use them and you don't have to keep a control connection
open on the database.
But there are so many ways to do it!. It always confuses
Yep no probs at all. You can have multiple databases open at the same
time and open and close them on different threads as you like.
You'll just need to do the appropriate connection handling linked to the
session open / close.
Scaling is going to depend on your application architecture too.
Maybe you are better off with a different database for each book, and
open and close them 1 at a time?
It really depends on your application...
On 1/06/2015 9:23 PM, naseh wrote:
version 1.4.187
i have designed the DB to have a table for every book. maybe it is not
a good design but it is
Isn't a bit unnerving if any client can actually corrupt the database?.
I thought it is the servers job to keep the database safe?
On 8/05/2015 7:10 PM, Rami Ojares wrote:
Hold your horses!
I found one more client that accesses the database with version 186.
I had forgotten all about it.
It
I think you have to turn off the space re-using or something like that
while running the backup?
Maybe that should happen automatically as part of the script tool?
On 8/05/2015 7:48 AM, Rami Ojares wrote:
I tested it out and it seems that the backup becomes completely
corrupted.
So one
when doing a backup?
Either this, or use the backup SQL statement. With the MVStore
storage, this should work very well. With the PageStore, it should
also work, but I didn't test as much.
Regards,
Thomas
On Friday, May 8, 2015, Ryan How r...@bitworks.com.au
mailto:r...@bitworks.com.au
Heya,
What's your connection string?
On 5/03/2015 8:16 AM, Bashorun Olajide wrote:
i am new to H2, i just started 2 days ago after reading about javafx. i got an
error while trying to retrieve record in a database. i dont even know maybe
successfully connection established.
i got this error
I'll let someone more qualified give you a better answer.
But, yes I can't see any reason why it wouldn't work (The same as if it
was on a read only filesystem?)
I find embedded mode has better performance, but it is more profound in
lots of smaller queries. If it is larger less frequent
and without locking, while
the first process opens it r/w and creates a lock file and would they
interfere with each other?
On 28 February 2015 at 17:43, Ryan How r...@bitworks.com.au
mailto:r...@bitworks.com.au wrote:
I'll let someone more qualified give you a better answer.
But, yes
I think with the MVStore it only works with READ_COMMITTED. Other
isolation levels don't work.
If you use the page store it should work (or the latest stable version).
On 28/02/2015 10:25 AM, Sachin Patil wrote:
Hello there,
I am trying use H2 database(1.4.185) in junit environment and
I use it on the 1.3 branch in production. It was considered experimental
in 1.3.
There are issues with blobs, I can't remember what it is, but we ended
up not using them. Apart from that we've had no issues in production
over the past 2 years with MVCC.
It may have been blobs and FOR
Just out of interest, something like an Access database works out of the
box on windows (I've never seen any glaring corruption issues, they
corrupt easily, but not as easily as a bug like this would indicate).
I wonder if it works around it? I'm sure older versions like Access 2003
are
Hi,
I don't know how it is actually implemented, but conceptually It depends
on your query.
If you do an order by and limit, it'll need to sort it all first before
pulling the first record. So it isn't going to shortcut the query at all.
The limit will improve performance because it only
I've had similar issues with anti-viruses locking the file while it
unzips it and scans it...
On 10/12/2014 1:03 AM, Thomas Mueller wrote:
Hi,
Could you describe what you do exactly (how you create the backup, how
you copy it)? How do you know the file is not available?
Regards,
Thomas
Is it CPU bound or disk bound during that 90 seconds when the query is
running?
You can try increasing JVM memory?
The the slow query the same as all the others?. perhaps it isn't using
an index for that query?
You really need to try and find out what is going on, or get it running
in a
I think it means that is why it doesn't sync for every commit, so it can
get higher performance than that.
If it syncs for every commit, then if you are committing each row and
doing a lot of updates for example, then you get bad performance.
Probably wouldn't matter much for an SSD, but on a
Pretty sure he is using the backup command.
Here's the code:
String sqlStatement = BACKUP TO ?;
try (Connection connection = getConnection())
{
try (PreparedStatement preparedStatement =
I have an application that uses BACKUP to copy the database to tablets
for offline use. It fails and produces a corrupt database about once a
year. It probably does 3000 or so copies a year. But like you, I'm not
sure if it is in the backup code or the copy code and I can't reproduce
it in a
It isn't safe to backup the file that way. You need to use the inbuilt
backup command if the database is in use.
http://h2database.com/html/grammar.html#backup
http://h2database.com/javadoc/org/h2/tools/Backup.html
On 3/09/2014 6:27 AM, Kam wrote:
I don't know what's causing it. We have a
I thought it did this already?
It just saves the string once and uses the hash or something?
I can't find it, but I thought I read such a thing in my years or forum
browsing on here... It was like an internal de-duplication that happened
by default. Maybe I was dreaming...
Ryan
On
H2 is not multi-threaded in this version. So a single long running
statement will block everything else. You'll either need to split it up
into smaller inserts (can still be a single transaction, just smaller
queries), or you could try turning on MULTI_THREADED=1.
There is multi-threaded
The 2nd digit incremented aswell from 1.3 to 1.4
On 4/05/2014 1:52 PM, Lukas Eder wrote:
Am Samstag, 3. Mai 2014 14:10:11 UTC+2 schrieb Noel Grandin:
On Sat, May 3, 2014 at 12:46 PM, Lukas Eder lukas...@gmail.com
javascript: wrote:
What does Beta mean when upgrading to any
Yeah hash comparison on writes is a bit of overkill. If you can't rely
on what is written to disk to actually be written to disk we are all in
major trouble!
Also, java handles exceptions for permissions, out of disk space, etc,
pretty well. just try, catch with appropriate handling as per
Hi,
I'd try using the latest 1.3 version. I've had it working from Access
and Excel for quite some time, but I don't use it regularly.
Ryan
On 23/04/2014 3:56 AM, Sylvain Nault wrote:
I can see the Support for the PG_CATALOG is part of the version 1.4.x
roadmap. Will it work with the MS
Hi,
I was just reading your previous posts. It may be possible you are
filling up your disk buffers after so many writes, so it is quite fast,
then slows down while the OS flushes to disk. So you probably find it is
fast up to a point, then slows right down after then. So if you try with
I'm really not sure sorry. I wouldn't think it would cause an issue if
it doesn't OOM. Even in the event of OOM or power outage, process
killed, etc, it still should be fairly safe.
Can you post about your environment, JVM version, OS, JVM startup params
etc? Do you access files on a network
Just another note, you aren't getting Out of Memory Errors are you?. The
docs say the Soft LRU cache is experimental and has been known to cause OOM?
On 3/04/2014 10:29 PM, Steve McLeod wrote:
Hi David,
I use H2 as an embedded database in desktop software with a
substantial user base. I
You need to enter a jdbc url, not the file path.
eg. jdbc:h2:file:c:\path\to\my\db
Check the docs for valid URLs for h2.
On 29/03/2014 1:42 PM, Axel wrote:
Hiya my name is Alex.
I am attempting to connect to a local DB but I get a error No suitable
driver found and a java error.
I have
Hi,
I'm not sure multithreaded is 100% safe yet. I wonder if it was caused
by an issue with that.
On 7/03/2014 2:06 PM, Alex wrote:
hi
db 1.3.172
FILE_LOCK=FS;PAGE_SIZE=1024;CACHE_SIZE=400072;MULTI_THREADED=TRUE;DATABASE_TO_UPPER=FALSE
i guess, some unclear way unique index has been
If it is a small script you can copy paste the script in I think.
If you want to run from a file then issue the RUNSCRIPT sql command. If
it is on a server, the script needs to be accessible on the server and
use a path as it would be on the server.
What connection string are you using? LOBs and MVCC are a bit stuffed...
On 15/02/2014 2:08 AM, Darren S wrote:
I seemed to hit the error Missing lob entry quite frequently for my
particular application. I was wondering if there are some settings I
should enable, or some information I can
Yeah it doesn't actually affect my use cases, as long as there are no
corruptions I'm happy. If I wanted to make sure everything is written to
disk straight away I'd probably manually make a kernel call to flush
system and disk caches and need to test it on each platform.
Some RAID
Hi,
What args are you actually putting into svrArgs? Perhaps on that machine
svrArgs is incorrect?
On 30/01/2014 7:18 AM, Subhash Agrawal wrote:
Hi All,
We are using H2 database in embedded mode and use
Server.createTcpServer(svrArgs).start() to get instance of h2 server.
It works fine
on this?
On Wed, Jan 29, 2014 at 4:55 PM, Ryan How r...@exemail.com.au
mailto:r...@exemail.com.au wrote:
Hi,
What args are you actually putting into svrArgs? Perhaps on that
machine svrArgs is incorrect?
On 30/01/2014 7:18 AM, Subhash Agrawal wrote:
Hi All,
We are using H2
On 28/01/2014 3:41 AM, forkj...@gmail.com wrote:
# Storing the content yourself on the filesystem and link the content to
a path column on a database table?
I do the latter in my project. But you lose any transactional
consistency unless you create a complex mechanism. So I'm sure I have a
On 17/01/2014 3:29 AM, Thomas Mueller wrote:
Hi,
from my experience I can propose to use SSDs for the storage.
Yes, that's a very good point. SSDs are much faster, specially for
reading, and for writing larger blocks.
SSD has much better random read and write, which is awesome for
I think Thomas would know this.
But my gut feeling is no. I think it uses Java 6, so I can't imagine it
using the Java 7 APIs.
But it does have pluggable filesystems, so you could probably write your
own Java 7 NIO2 plugin, then use the java 7 filesystems under that. But
I have no
If the timestamps are unique you could try including a subquery which
gets the next lowest date
eg. (not sure if this is valid SQL syntax, I'm not meant to be on the PC
:) Select data, datetimefield, (select top 1 datetimefield from table1
where datetimefield A.datetimefield order by
Wouldn't you usually do that in your reporting software? Or on any layer
on top of the database?
In MS Access I've done it using a custom function which stores the last
value and then only outputs the next value if it is different. Not sure
how to do that in H2 ? Probably via a custom
I've been using it in a production environment for several years without
issue. But it isn't very high concurrency, so I've probably just been
lucky that it hasn't had issues.
On 2/01/2014 4:39 PM, Uli wrote:
Hi Noel,
what a pity! Thanks for digging into the synching problems!
It sounds to
On 2/01/2014 6:40 PM, Thomas Mueller wrote:
Would be nice if there was some java in-memory file abstraction
I don't understand, you can already use the file system abstraction of
H2. You can also use databases that persist to an in-memory file
system, using the database URL
[] script = IOUtils.readBytesAndClose(in, -1);
System.out.println(new String(script, UTF-8));
}
}
Or you can just run script and process the result set.
Regards,
Thomas
On Thu, Jan 2, 2014 at 12:14 PM, Ryan How r...@exemail.com.au
mailto:r...@exemail.com.au wrote:
On 2/01/2014
On Thu, Jan 2, 2014 at 12:14 PM, Ryan How rh...@exemail.com.au
javascript: wrote:
On 2/01/2014 6:40 PM, Thomas Mueller wrote:
Would be nice if there was some java in-memory file abstraction
I don't understand, you can already use the file system
abstraction
On 2/01/2014 12:43 PM, Rob Oaks wrote:
Thank you both for your assistance.
Ryan, it’s comforting to know that someone has accomplished something
similar, which inclines me to go ahead and build a prototype. I have
just a couple more questions:
1.
I would prefer to avoid
It will be 2. and the second create table will fail because the table
already exists. But I don't see how that has anything to do with being
thread safe? The in memory database is per process, so all threads
access the same database. If you want a separate database per thread
then give the
Are you possibly copying the database while it is still open on system A?
On 11/12/2013 12:47 AM, Johann Petrak wrote:
I created a database (two files) on system A, then moved the files to
system B and when I want to connect to that DB on system B, the first
time just establishing the
I use it with H2 every day. Works great.
You are quite possibly pointing to an empty database as suggested...
Here is my jooq.xml file I use to generate the classes. Note that
because the h2 database path is relative, I need to have the working
directory set correctly.
?xml version=1.0
Are you using an in memory database? They don't persist so will be blank
each time you open it.
On 3/12/2013 11:29 AM, ??? wrote:
Do not know why..
11:17:41,719 INFO ~ Connected to
jdbc:h2:tcp://10.1.1.222/~/test;CACHE_SIZE=131
072;CACHE_TYPE=SOFT_LRU;WRITE_DELAY=2000
11:17:41,747 INFO
Not sure if I would choose H2 for a 2TB database. I don't think it has
been tested too well with databases of that size. It would probably need
a lot of RAM.
I'd be inclined to try smaller files rather than a single 2TB file. For
instance backing up, or recovery of corrupt databases will be
No idea. Must be an issue in your config. Try opening another connection
to your DB from the H2 console and see what happens to the tables.
On 3/12/2013 1:32 PM, 董藝興 wrote:
NO,i setting file application.conf is connection
http://www.h2database.com/html/features.html#embedded_databases
database
This is interesting :)
Does MVCC even support different isolation modes?. I thought it was it's
own isolation mode? The different isolation modes don't seem to make
sense in an MVCC context?
Thanks, Ryan
On 24/11/2013 11:50 PM, cowwoc wrote:
Hi Thomas,
It was my understanding that
possible case and get that
working first. I have had other problems with deleting files from java
on windows 7.
- mike
On Mon, Nov 11, 2013 at 5:16 AM, Ryan How r...@exemail.com.au
mailto:r...@exemail.com.au wrote:
The database will close when all connections to it are closed. So
you
The database will close when all connections to it are closed. So you
must still have some connections open (Are you using a connection pool??)
Also see http://h2database.com/html/grammar.html#set_db_close_delay
Is there still the lock file present?. If that is still present then
there is an
No Worries :). If you run Explain Analyze update tre_combinedtestday set
compoundpercent=-13.373487168551002, totalpercent=-5.466540999057416,
dollaramount=0.9453345900094259 where id=230827 what do you get?
On 29/10/2013 2:12 PM, TrendTimer.com wrote:
Kartweel, I forgot to say thanks for
Have you got an index on the ID column?, it doesn't appear to be using
an index. Usually you would have ID as the primary key which would
create an index...
On 29/10/2013 3:17 PM, TrendTimer.com wrote:
Explain Analyze update tre_combinedtestday set
compoundpercent=-13.373487168551002,
Can you type in the dir command and show the directory listing, then
the java -cp h2*.jar etc command and copy paste the output here?.
On 24/10/2013 5:58 PM, sheing3003 wrote:
Hello All,
When I tried to do jave -cp h2*.jar org.h2.tools.Server (or any
others like Shell), windows command
I you can paste the entire output and directory listing, I'm not sure
what you are doing. It seems like a basic java classpath issue. Nothing
H2 related.
On 25/10/2013 12:17 AM, sheing3003 wrote:
Could not find the main class: org.h2.tools.Server in
I quite often use a huge IN clause as a quick way to combine lots of
small queries. I'm fairly sure it uses the index (although I must admit
I haven't checked). But it usually speeds up the queries considerably.
Much faster than lots of small queries. Even in embedded mode for me.
On
The larger the database the more RAM you need generally. It depends what
you mean by large. I have databases of a few GB that run without very
high memory usage, but if you start running large transactions it
requires more memory.
So it is probably more how you use the database rather than
.
Regards,
Thomas
On Tue, Sep 10, 2013 at 10:51 AM, Ryan How
rh...@exemail.com.au wrote:
Howcome it is getting OOM on import? It shouldn't do
that should it? I've imported several GB of data
before, I did
Howcome it is getting OOM on import? It shouldn't do that should it?
I've imported several GB of data before, I did get OOM, but increasing
heap size to around 1GB worked for me. I didn't need to go to crazy sizes...
On 10/09/2013 4:32 PM, Noel Grandin wrote:
On 2013-09-06 20:15, Brian Craft
On 6/09/2013 11:48 PM, Daniel Stone wrote:
We did also experience many problems with a “missing lob” error which
actually triggered the cluster issue in the first place. We have
managed to work around this but is this still a bug with H2 (I have
seen this problem mentioned on the web)?
It sounds like your configuration is corrupted. It cannot find the H2
driver. Just make sure you still have the H2 driver on the classpath.
On 4/09/2013 9:47 PM, James Ball wrote:
I have been using Jubula to programme test cases and during the middle
of one Jubula froze and would not respond.
works for me.
you want IS NOT NULL, IS NULL instead of = null
This is what I did
SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 'FILE_RESOURCE'
SELECT * FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 'FILE_RESOURCE';
Sounds like you're on the right path. It's usually good practice to have
a connection per thread. A H2 connection is thread safe, but I don't
think that is the case for all DBs. But anyway, a connection per thread
is better, because then if you turn on multithreaded mode for H2 you
will get
On 21/07/2013 1:03 AM, Thomas Mueller wrote:
But still, there is a risk. This was one reason for me to start
working on the new storage engine MVStore: to reduce the risk of
corruption (the MVStore is quite a bit simpler than the current,
traditional storage engine). It is still work in
If you look inside the zip file of your backup are there trace files in
there?
On 18/07/2013 4:14 PM, EDV-Services wrote:
Hi Thomas,
I thought the same, I thought it could be a bug in my application.
Then I restored my latest backup file by using the web interface of H2
by clicking on Tools
I shut it down about 50 times a day uncleanly during development and
never had an issue. As the database gets larger startup time takes
longer. I guess as it checks the database. It all happens automatically.
I can't vouch for ext4, I've had a lot of corruptions on that (not H2,
but just the
Hi,
Yeah this doesn't work. And nobody is currently supporting MVCC until
the new Storage engine is out. So options are: Don't use MVCC. Don't use
For Update with LOBS. Try and fix H2 code.
Also, Other people have noted problems with lots of LOBS and MVCC when
not using for update.
Ryan
But it does re-use space. So if you delete data, then insert new data,
the database won't grow, it will overwrite the deleted data.
Please correct me if I am wrong.
On 14/06/2013 2:33 PM, Noel Grandin wrote:
You can use DISK_SPACE_USED to get the approx amount of space used by
a table
By Default H2 run single threaded, so when a statement is executing, it
blocks everything else. See
http://h2database.com/html/features.html#multiple_connections
You can turn on multithreaded mode if you really need the concurrency,
or have single long running statements and don't want to
The stack trace indicates version 1.3.170 ?
On 3/06/2013 5:51 AM, Martin Lichtin wrote:
We are using 1.3.172 and have seen a very similar NPE.
2013-06-01 14:48:14,035 [Atomikos:1 ] WARN XAResourceTransaction - XA
resource 'JDBC-H2RuleExecCamelITest': rollback for XID
Have you got the pgserver running in H2 ? on Port 5432?
From memory the default Postgres port default H2 pgserver port might
be different...
On 3/06/2013 10:41 AM, Balaji M wrote:
Hi,
When i try to connect through DSN less connection i am getting
error pls help
/Could not Connect to
Sorry, so do you mean you are still getting a very similar stack in
1.3.172?, You just don't have a record of it?
On 3/06/2013 3:12 PM, Martin Lichtin wrote:
Yeah, you're right. I already updated the version, but the stack
indeed came from 1.3.170.
On Monday, June 3, 2013 8:58:13 AM UTC+2,
When you start the PGServer you can also set the port it is listening on.
[-pgPort port]The port (default: 5435)
On 3/06/2013 3:55 PM, Balaji M wrote:
Hi ,
I am new to the H2 ...
For ODBC DSN connection by default the port no is 5435 ..it is working
fine
For ODBC DSN less
I imagine you could probably change the source of the shell tool fairly
easily to change the max length, or make it configurable as a command
line option?
And also make header output optional.
Or you could probably strip the header by piping it through some awk
command or something (I'm a
I'm using MVCC and the only problems I've come across appear to be with
LOBS (that's ok for me, don't really need them for my use case).
What I mean is that it may have been less work for you to try and fix
the H2 source (Seeing the issues seem to all be with LOBS) rather than
change to
I'd try the connection pooling :). I imagine that means it keeps a pool
of connections open to the database underneath, then when it needs a
connection in the GUI there is always one available.
I'd imagine the connection string would be the same, connection pooling
would be handled by
Hi,
You can't currently do that in H2. You need to do it in the firewall on
the machine.
Ryan
On 26/05/2013 12:33 PM, Jay Lee wrote:
Hi All,
I am using H2DB in TCP server mode.
The only thing missing or I can't find is that
how to block clinet connection based on IP address.
For example
I'm seeing 171 on the 2nd line and 169 on the 4th line?
On 24/05/2013 11:25 PM, Thomas Mueller wrote:
Hi,
The error message Caused by: java.lang.NoClassDefFoundError:
org/h2/engine/Constants doesn't say much about the state of the
database file. All it says is that you have a classloading
I've been using MVCC for quite a while. It appears pretty stable, but
there is a bug with FOR UPDATE and LOBS. (It doesn't work), so there are
still bugs. Furthermore MVCC is not really supported by anyone, so if
you come across bugs you'll probably need to fix them yourself.
I think once the
Hi,
Wouldn't that indicate each value is in a new column, being a CSV file?
I'd imagine in your export process to get the CSV you'd need to not
export the commas ?
Ryan
On 23/05/2013 3:30 AM, pkod...@uic.edu wrote:
Hi,
I would like to know if there is any in built function to remove
I don't know of anything inbuilt. You could dump it into a varchar
column, remove all the commas (SET field = REPLACE(field, ' ')), then
alter column type to decimal?
Ryan
On 23/05/2013 3:41 AM, Rnj wrote:
I'm trying to import a csv file into a h2 database table. I have a
number 786,960,000
Hi,
That is how the identity columns work. Also if a failed transaction then
they will still increment. If the database is not closed correctly they
may jump by a large number. There is no guarantee that they will be in
increments of 1, just that they will be higher in value than the last.
Well you could get a global database lock with every transaction, it
would kill concurrency, but you'd be ensured to never get a deadlock :P
Sorry, not very helpful :), but no-one else has replied on here :)
On 18/05/2013 7:58 AM, Gili wrote:
Hi,
I wanted to get your opinion
regarding
I'm not sure how dropbox handles open files, but regarding multiple
files h2 should be fine as everything is in 1 file. If you have separate
lobs or fulltext index that might be a problem. You could regenerate
full text index, but if lobs are out of sync I don't know how it would
handle it. So
Can you paste a stack trace of what error you are getting?
And what URL are you using to connect to the database?
Thanks, Ryan
On 8/05/2013 1:54 PM, khang phamngoc wrote:
I used H2DB for my project, well evrything seem good but 2 days ago
when I'm using it, my computer shut down so when I
Dropbox just syncs a folder on the local PC.
So it depends on what you want to do with it?
If you want to use it on more than 1 PC at a time it won't work (You'll
just get conflicts). Otherwise it should work ok I think. ? I'm not sure
how dropbox accesses in use files. But it might only sync
Sounds like you are running out of memory.
What is the memory usage of the java process? and what JVM parameters
are you using?
You'd expect CPU to be maxed out for an in-memory database. Something
would have to max out.
On 1/05/2013 9:40 PM, neelima kaliki wrote:
Below are my System
You can try it and see if it works.-Xmx7000m or something like that. But
you may keep running out of memory. It might just be too much data to
fit. You might be better off trying using a file based database on an SSD?
On 2/05/2013 12:03 AM, neelima kaliki wrote:
So, if i increase my memory
Strange that it took 10 minutes to reproduce it? That would almost be
consistent with the sub-optimally executed view?
I'm wondering whether it might be a caching thing?
It's hard without a reproducible test case. I'm just guessing what it
might be and trying to narrow down the
Wow, that isn't a simple query is it :)
Nothing seems obvious does it.
And if you re-create the view then you get the same execution plan as
running the select directly?
I think you'll need someone much more knowledgeable than me helping.
Hope you get it sorted! I'd only be trying different
1 - 100 of 250 matches
Mail list logo
