[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
This is the copy of my post: I confirm that this is a change in behaviour between 3.2.3 and 3.2.4RC2. 3.2.3 has a bug in on-find read-ahead for single-object finders. I.e. when the result of a finder is a single object which is preloaded on-find with some load group which doesn't cover all the fields in the default load group (*), first access (even to the field that was preloaded) to this found instance will load the default load group anyway. Which is obviously wrong. The difference between 3.2.3 and 3.2.4 is that if at least some data was loaded from read-ahead cache 3.2.4 does not proceed to load the default load group while 3.2.3 does. So, in the case above, the data (CMR field) was preloaded and other CMP field access triggers lazy-loading. This is expected behaviour according to the fix. I still have to think about this. To prevent this you can: 1) define a lazy-load group and load all the fields at once on first access with on-load read-ahead (like 3.2.3 in this case) 2) define on-find read-ahead for the relationship and preload all the data at once at CMR access. This is even better than 1). View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3836390#3836390 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3836390 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
No you were not dreaming, or we were all dreaming. But it seems that the server falls down many times actually. Its not Alex who tells you the "on-find" strategy but me. But Alex post after me saying that you should keep your on-load strategy, I don't know why. on-load load the beans one by one on-find load a page size each time on bean is not in current page Do you remove some options in your datasource file (pool size or something like this) ? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3836330#3836330 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3836330 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
Was I dreaming last night ? I think not. I'm sure I saw a posting from alex loubiansky yesterday evening here. He told me to put "read-ahead="on-find"" onto the relation. I tried this with success. Now again only one DB-Access in server.log for reading the all the rows with all the fields from the related table. Thanks a lot. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3836329#3836329 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3836329 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
Hi, Before the crash, We said in this topic that lafr was using the 3.2.4 version, and I said that it was working fine in the 3.2.3 version. So, which version has most bugs ? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835749#3835749 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835749 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
I confirm that this is a change in behaviour between 3.2.3 and 3.2.4RC2. 3.2.3 has a bug in on-find read-ahead for single-object finders. I.e. when the result of a finder is a single object which is preloaded on-find with some load group which doesn't cover all the fields in the default load group (*), first access (even to the field that was preloaded) to this found instance will load the default load group anyway. Which is obviously wrong. The difference between 3.2.3 and 3.2.4 is that if at least some data was loaded from read-ahead cache 3.2.4 does not proceed to load the default load group while 3.2.3 does. So, in the case above, the data (CMR field) was preloaded and other CMP field access triggers lazy-loading. This is expected behaviour according to the fix. I still have to think about this. To prevent this you can: 1) define a lazy-load group and load all the fields at once on first access with on-load read-ahead (like 3.2.3 in this case) 2) define on-find read-ahead for the relationship and preload all the data at once at CMR access. This is even better than 1). View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835696#3835696 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835696 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
The session beans with "Supports" are not relevant here. Although I switched tags from "Supports" to "Required" for all beans. The EntityBeans are called from a JSP via session bean CustomerOrder (which already has tag "Required"). I tried 3.2.3 today with my ear and here I got the expected behaviour. The problem occurs with also with the first access. So I think this is really a bug in 3.2.4RC2. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835892#3835892 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835892 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
It could be a bug, however, some of your front end sessions are running with "Supports" transaction tag, so make sure you have transaction associated with the thread when you're accessing CMP entities. Try first with "Required" tag on all sessions. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835851#3835851 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835851 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
Your descriptors are OK, means using default values. Do you try on JBoss 3.2.3 ? I have seen many peculiarities for the ejb plugin on JBoss 3.2.4 series. Perhaps try also to optimize loading. Is your problem occurs after a long run or from the start ? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835768#3835768 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835768 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
I think I don't play with eager and lazy loading. I also have seen it working some time ago. Current I use JBoss 3.2.4RC2. Deployment descriptors are generated by XDoclet 1.2.1 and are accessible at http://www.lafr.de/ejb-jar.xml and http://www.lafr.de/jbosscmp-jdbc.xml. The EB Afsta ist called from SLSB CustomerOrder. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835724#3835724 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835724 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
anonymous wrote : Is this perhaps a bug ? Why is everyone so obsessed with blaming any problem on a bug at the moment? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835614#3835614 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835614 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
Hi, It looks like you are playing with eager and lazy loading options. As far I as know, nobody here has seen this problem. I use myself JBoss 3.2.3 and all the fields are loaded in one shot. Give us more details about the JBoss version and perhaps your descriptors. Pascal View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835600#3835600 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835600 --- This SF.Net email is sponsored by: Oracle 10g Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE. http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
This is not a bug. It is the way JBoss is doing it. Please see, http://www.jboss.org/index.html?module=bb&op=viewtopic&t=49517 /Goran View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835590#3835590 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835590 --- This SF.Net email is sponsored by: SourceForge.net Broadband Sign-up now for SourceForge Broadband and get the fastest 6.0/768 connection for only $19.95/mo for the first 3 months! http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence & CMP/JBoss] - Re: curious SQL-Statements when loading related data
I think you misunderstood the problem description. My problem is not the where clause of the statement. The problem is, that the first access gets the PK-field plus first additional field, the second access reads the same rows, but know the PK-field plus second additional field, and so on and so on. I expected 5 select to read the 84 rows with max 20 at a time (4x20+1x4). Instead I get 315 selects (factor 63 because of 63 additional columns in afpos). View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3835595#3835595 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3835595 --- This SF.Net email is sponsored by: SourceForge.net Broadband Sign-up now for SourceForge Broadband and get the fastest 6.0/768 connection for only $19.95/mo for the first 3 months! http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
