Re: 9.2.0.4 anyone
Hi! Which platform is it? I checked on one 4CPU Linux server with 160MB shared pool it defaulted to 1. Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Saturday, September 13, 2003 1:49 AM On the 9.2.0.3 databases I checked, it is 3. -Original Message- Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Tanel Poder INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Tru64 5.1A. 12 CPU box. Two different instances on the box. -Original Message- Hi! Which platform is it? I checked on one 4CPU Linux server with 160MB shared pool it defaulted to 1. Tanel. -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
How can I not see the init parameter, _kghdsidx_count in 9.2.0.3.0? From: Tanel Poder [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Subject: Re: 9.2.0.4 anyone Date: Thu, 11 Sep 2003 17:54:30 -0800 Hi! Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning causing shared pool latch contention. Various experts suggesting to increase your shared pool make the situation even worse, because in time, the LRU and freelists get even longer, thus one scan takes more time to complete or fail. _kghsidx_count = 1 (The correct name is _kghdsidx_count) So, in 9i Oracle introduced the ability to create several heaps for serving shared pool. If you've set the shared_pool_size to 64M and _kghdsidx_count to 4, you'll have 4*16M shared pool heaps, every heap having its own descriptor, extents, LRU lists, freelists and shared pool child latch (there is no such child latch in 8i). Shorter lists mean faster scanning, more latches mean more scalability when serializing access to a resource. The downside is, that usually the resource has to be split that every latch protects one and only one part of resource. Thus if the shared pool is split into 4 parts and all memory allocation requests happen to use the same freelist for some reason, only 25% of memory can be used. This is the reason why you can avoid ORA-4031s when setting _kghsidx_count to 1 - this practically enables the old behaviour. In 8i there is only one set of freelists, in 9i there can be more sets, default is 1 and max limit is 7 as far as simple testing on my 9.2.0.4/W2k has showed - I have only 7 shared pool child latches and didn't find a way to increase them. When you increase _kghdsidx_count, you see more lines in x$kghlu as well, one for each heap (normally there was only one). Also you can verify the behaviour when taking SGA heapdump at level 2 and search for HEAP DUMP or FREE LISTS in trace, there are as many free lists in dump, as you've stated with _kghdsidx_count init parameter. (there is one extra heap in dump, this is the SGA parent heap for shared pool heaps). that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line
RE: 9.2.0.4 anyone
Is this for all platforms? I haven't heard about it for AIX5.2 but that doesn't mean I shouldn't apply it. Thanks in advance, Ruth -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Stephen Lee Sent: Thursday, September 11, 2003 5:29 PM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) _kghsidx_count = 1 that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Ruth Gramolini INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
It's a hidden parameter, use this query instead: select n.ksppinm name, v.ksppstvl value from x$ksppi n, x$ksppsv v where n.indx = v.indx and n.ksppinm = '_kghdsidx_count'; Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Friday, September 12, 2003 5:59 PM How can I not see the init parameter, _kghdsidx_count in 9.2.0.3.0? From: Tanel Poder [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Subject: Re: 9.2.0.4 anyone Date: Thu, 11 Sep 2003 17:54:30 -0800 Hi! Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning causing shared pool latch contention. Various experts suggesting to increase your shared pool make the situation even worse, because in time, the LRU and freelists get even longer, thus one scan takes more time to complete or fail. _kghsidx_count = 1 (The correct name is _kghdsidx_count) So, in 9i Oracle introduced the ability to create several heaps for serving shared pool. If you've set the shared_pool_size to 64M and _kghdsidx_count to 4, you'll have 4*16M shared pool heaps, every heap having its own descriptor, extents, LRU lists, freelists and shared pool child latch (there is no such child latch in 8i). Shorter lists mean faster scanning, more latches mean more scalability when serializing access to a resource. The downside is, that usually the resource has to be split that every latch protects one and only one part of resource. Thus if the shared pool is split into 4 parts and all memory allocation requests happen to use the same freelist for some reason, only 25% of memory can be used. This is the reason why you can avoid ORA-4031s when setting _kghsidx_count to 1 - this practically enables the old behaviour. In 8i there is only one set of freelists, in 9i there can be more sets, default is 1 and max limit is 7 as far as simple testing on my 9.2.0.4/W2k has showed - I have only 7 shared pool child latches and didn't find a way to increase them. When you increase _kghdsidx_count, you see more lines in x$kghlu as well, one for each heap (normally there was only one). Also you can verify the behaviour when taking SGA heapdump at level 2 and search for HEAP DUMP or FREE LISTS in trace, there are as many free lists in dump, as you've stated with _kghdsidx_count init parameter. (there is one extra heap in dump, this is the SGA parent heap for shared pool heaps). that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ
Re: 9.2.0.4 anyone
Thanks, Tanel. You always provide good solution. From: Tanel Poder [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Subject: Re: 9.2.0.4 anyone Date: Fri, 12 Sep 2003 07:14:24 -0800 It's a hidden parameter, use this query instead: select n.ksppinm name, v.ksppstvl value from x$ksppi n, x$ksppsv v where n.indx = v.indx and n.ksppinm = '_kghdsidx_count'; Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Friday, September 12, 2003 5:59 PM How can I not see the init parameter, _kghdsidx_count in 9.2.0.3.0? From: Tanel Poder [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Subject: Re: 9.2.0.4 anyone Date: Thu, 11 Sep 2003 17:54:30 -0800 Hi! Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning causing shared pool latch contention. Various experts suggesting to increase your shared pool make the situation even worse, because in time, the LRU and freelists get even longer, thus one scan takes more time to complete or fail. _kghsidx_count = 1 (The correct name is _kghdsidx_count) So, in 9i Oracle introduced the ability to create several heaps for serving shared pool. If you've set the shared_pool_size to 64M and _kghdsidx_count to 4, you'll have 4*16M shared pool heaps, every heap having its own descriptor, extents, LRU lists, freelists and shared pool child latch (there is no such child latch in 8i). Shorter lists mean faster scanning, more latches mean more scalability when serializing access to a resource. The downside is, that usually the resource has to be split that every latch protects one and only one part of resource. Thus if the shared pool is split into 4 parts and all memory allocation requests happen to use the same freelist for some reason, only 25% of memory can be used. This is the reason why you can avoid ORA-4031s when setting _kghsidx_count to 1 - this practically enables the old behaviour. In 8i there is only one set of freelists, in 9i there can be more sets, default is 1 and max limit is 7 as far as simple testing on my 9.2.0.4/W2k has showed - I have only 7 shared pool child latches and didn't find a way to increase them. When you increase _kghdsidx_count, you see more lines in x$kghlu as well, one for each heap (normally there was only one). Also you can verify the behaviour when taking SGA heapdump at level 2 and search for HEAP DUMP or FREE LISTS in trace, there are as many free lists in dump, as you've stated with _kghdsidx_count init parameter. (there is one extra heap in dump, this is the SGA parent heap for shared pool heaps). that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance
Re: 9.2.0.4 anyone
I don't recommend setting any hidden _* parameter just because someone thinks it's neat or it has once worked form someone. If you got a problem, which isn't solvable by any other mean, then turn to underscore parameters or other unsupported actions... Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Friday, September 12, 2003 6:04 PM Is this for all platforms? I haven't heard about it for AIX5.2 but that doesn't mean I shouldn't apply it. Thanks in advance, Ruth -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Stephen Lee Sent: Thursday, September 11, 2003 5:29 PM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) _kghsidx_count = 1 that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Ruth Gramolini INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Tanel Poder INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
There is a known bug in Oracle's handling of the shared pool in 9.2.0.X (at least) where the shared pool becomes increasingly fragmented and eventually gets to where no more space can be allocated. Attempts to flush the shared pool to clear it fail (Yes, I know this means everything has to be re-parsed) and the instance must be restarted. Even attempts to flush the shared pool at regular intervals as a preventative measure fail. The latest TAR we had on this proved to be another worthless endeavor with me going on hours of wild goose chases with RDA reports, stats pak reports, trace files, and e-mail writing, all the while I keep trying to get them to acknowledge this is a known bug; and does Oracle have any known useful work arounds for it; and has Oracle determined a release in which the bug is fixed; only, end the end, to have them recommend two things that *I* told them: _KGHDSIDX_COUNT=1 (don't know how I could have possibly misspelled it earlier; it's such a natural name) and apply the 9.2.0.4 patch. Plus they decided to go off on some tangent about performance tuning when performance tuning was not the issue. You know, when you have this fatal 4031 thing constantly hanging over your head like some Sword of Damocles, you really would like to concentrate on that right now. I especially like it when they tell you to try this or try that in your production database to see what happens. Then when you continue to press the issue about this being a bug and insist that rather than banging around in the production database to see what happens, we need to know what work around Oracle has for this bug, they accuse you of not working with them to diagnose the problem with YOUR database. Hell! It isn't a problem with the database! It's a goddam bug in the Oracle code! What's the work around? When is it fixed? Besides, field services has been out here and looked at our configuration twice already. It's difficult to get them to understand that, just because there is some SQL that isn't using bind variables (Sorry, development is under a different management hierarchy), that shouldn't put the instance into a completely hosed and unusable state. Let's see now, we're talking about a product that claims it is worth $40,000 per CPU. I don't think MS Access has this problem. Whatever happened to Larry's Unbreakable label? For what it's worth, we were told by field services that the _KGHDSIDX_COUNT=1 essentially makes the shared pool into an 8.1.7 shared pool. Don't know how accurate that is; don't care; just want to know how to make the problem go away. Attempts to find out if any of this has been fixed in 9.2.0.4 were never answered, but the fact that 9.2.0.4 makes the _KGHDSIDX_COUNT=1 gives one the feeling Oracle chose to sweep the bug under the rug for now. Any useful suggestions on this would be most welcomed. While I'm at it here ... I mentioned in a previous post that it looks like the lock-ups on index builds appears to be fixed in 9.2.0.4. I recall there have been some posts about unexplainable, weird locking issues that nobody could answer. I wonder if maybe this is all related. -Original Message- Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning causing shared pool latch contention. Various experts suggesting to increase your shared pool make the situation even worse, because in time, the LRU and freelists get even longer, thus one scan takes more time to complete or fail. _kghsidx_count = 1 (The correct name is _kghdsidx_count) -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services -
RE: 9.2.0.4 anyone
Stephen, Surely, you must be mistaken. I just finished reading the Oracle 10G article in the latest Oracle magazine about how they regression test every little change made to the Oracle Rdbms code to ensure that bugs do not make it out into production. It *can't be* their fault. It *has to be* your database. Lets run some regression testing on your production instance - screw your users - to prove that it's your sql, not their code. Ain't life grand? :) Tom Mercadante Oracle Certified Professional -Original Message- Sent: Friday, September 12, 2003 11:39 AM To: Multiple recipients of list ORACLE-L There is a known bug in Oracle's handling of the shared pool in 9.2.0.X (at least) where the shared pool becomes increasingly fragmented and eventually gets to where no more space can be allocated. Attempts to flush the shared pool to clear it fail (Yes, I know this means everything has to be re-parsed) and the instance must be restarted. Even attempts to flush the shared pool at regular intervals as a preventative measure fail. The latest TAR we had on this proved to be another worthless endeavor with me going on hours of wild goose chases with RDA reports, stats pak reports, trace files, and e-mail writing, all the while I keep trying to get them to acknowledge this is a known bug; and does Oracle have any known useful work arounds for it; and has Oracle determined a release in which the bug is fixed; only, end the end, to have them recommend two things that *I* told them: _KGHDSIDX_COUNT=1 (don't know how I could have possibly misspelled it earlier; it's such a natural name) and apply the 9.2.0.4 patch. Plus they decided to go off on some tangent about performance tuning when performance tuning was not the issue. You know, when you have this fatal 4031 thing constantly hanging over your head like some Sword of Damocles, you really would like to concentrate on that right now. I especially like it when they tell you to try this or try that in your production database to see what happens. Then when you continue to press the issue about this being a bug and insist that rather than banging around in the production database to see what happens, we need to know what work around Oracle has for this bug, they accuse you of not working with them to diagnose the problem with YOUR database. Hell! It isn't a problem with the database! It's a goddam bug in the Oracle code! What's the work around? When is it fixed? Besides, field services has been out here and looked at our configuration twice already. It's difficult to get them to understand that, just because there is some SQL that isn't using bind variables (Sorry, development is under a different management hierarchy), that shouldn't put the instance into a completely hosed and unusable state. Let's see now, we're talking about a product that claims it is worth $40,000 per CPU. I don't think MS Access has this problem. Whatever happened to Larry's Unbreakable label? For what it's worth, we were told by field services that the _KGHDSIDX_COUNT=1 essentially makes the shared pool into an 8.1.7 shared pool. Don't know how accurate that is; don't care; just want to know how to make the problem go away. Attempts to find out if any of this has been fixed in 9.2.0.4 were never answered, but the fact that 9.2.0.4 makes the _KGHDSIDX_COUNT=1 gives one the feeling Oracle chose to sweep the bug under the rug for now. Any useful suggestions on this would be most welcomed. While I'm at it here ... I mentioned in a previous post that it looks like the lock-ups on index builds appears to be fixed in 9.2.0.4. I recall there have been some posts about unexplainable, weird locking issues that nobody could answer. I wonder if maybe this is all related. -Original Message- Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning
RE: 9.2.0.4 anyone
And for you cobol users out there: we have been told that the bug related to setting cursor_sharing=force when using cobol has been fixed in 9.2.0.4. So maybe that will help keep the 4031 bug in its box. -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
On the 9.2.0.3 databases I checked, it is 3. -Original Message- Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Where? I can only see 9.2.0.2 on the otn download site.. -Original Message-From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]On Behalf Of [EMAIL PROTECTED]Sent: 11 September 2003 01:20To: Multiple recipients of list ORACLE-LSubject: Re: 9.2.0.4 anyoneYeah, been out for some time now actually. [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09/10/2003 04:25 PM Please respond to ORACLE-L To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] cc: Subject:Re: 9.2.0.4 anyoneWhat? 9.2.0.4 is out now?RWBReginald W. BaileyIBM Global Services - ETS SW GDSD - Database ManagementYour Friendly Neighborhood DBA713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager)[EMAIL PROTECTED][EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED]Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet?Robert--Please see the official ORACLE-L FAQ: http://www.orafaq.net--Author: Freeman Robert - ILINET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L(or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).-- Please see the official ORACLE-L FAQ: http://www.orafaq.net-- Author: INET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
It's available as a patch from Metalink. Cheers, Mike -Original Message- Sent: 11 September 2003 10:59 To: Multiple recipients of list ORACLE-L Where? I can only see 9.2.0.2 on the otn download site.. -Original Message- [EMAIL PROTECTED] Sent: 11 September 2003 01:20 To: Multiple recipients of list ORACLE-L Yeah, been out for some time now actually. [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09/10/2003 04:25 PM Please respond to ORACLE-L To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED] cc: Subject:Re: 9.2.0.4 anyone What? 9.2.0.4 is out now? RWB Reginald W. Bailey IBM Global Services - ETS SW GDSD - Database Management Your Friendly Neighborhood DBA 713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager) [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED] Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert E mail Disclaimer You agree that you have read and understood this disclaimer and you agree to be bound by its terms. The information contained in this e-mail and any files transmitted with it (if any) are confidential and intended for the addressee only. If you have received this e-mail in error please notify the originator. This e-mail and any attachments have been scanned for certain viruses prior to sending but CE Electric UK Funding Company nor any of its associated companies from whom this e-mail originates shall be liable for any losses as a result of any viruses being passed on. No warranty of any kind is given in respect of any information contained in this e-mail and you should be aware that that it might be incomplete, out of date or incorrect. It is therefore essential that you verify all such information with us before placing any reliance upon it. CE Electric UK Funding Company Lloyds Court 78 Grey Street Newcastle upon Tyne NE1 6AF Registered in England and Wales: Number 3476201 -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Hately, Mike (LogicaCMG) INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
9.2.0.4 is a patchset - go to Patches section in metalink. Tanel. - Original Message - From: Mark Leith To: Multiple recipients of list ORACLE-L Sent: Thursday, September 11, 2003 1:59 PM Subject: RE: 9.2.0.4 anyone Where? I can only see 9.2.0.2 on the otn download site.. -Original Message-From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]On Behalf Of [EMAIL PROTECTED]Sent: 11 September 2003 01:20To: Multiple recipients of list ORACLE-LSubject: Re: 9.2.0.4 anyoneYeah, been out for some time now actually. [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09/10/2003 04:25 PM Please respond to ORACLE-L To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED] cc: Subject:Re: 9.2.0.4 anyoneWhat? 9.2.0.4 is out now?RWBReginald W. BaileyIBM Global Services - ETS SW GDSD - Database ManagementYour Friendly Neighborhood DBA713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager)[EMAIL PROTECTED][EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED]Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet?Robert--Please see the official ORACLE-L FAQ: http://www.orafaq.net--Author: Freeman Robert - ILINET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L(or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).-- Please see the official ORACLE-L FAQ: http://www.orafaq.net-- Author: INET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California-- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Where? I can only see 9.2.0.2 on the otn download site.. -Original Message-From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]On Behalf Of [EMAIL PROTECTED]Sent: 11 September 2003 01:20To: Multiple recipients of list ORACLE-LSubject: Re: 9.2.0.4 anyoneYeah, been out for some time now actually. [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09/10/2003 04:25 PM Please respond to ORACLE-L To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] cc: Subject:Re: 9.2.0.4 anyoneWhat? 9.2.0.4 is out now?RWBReginald W. BaileyIBM Global Services - ETS SW GDSD - Database ManagementYour Friendly Neighborhood DBA713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager)[EMAIL PROTECTED][EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED]Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet?Robert--Please see the official ORACLE-L FAQ: http://www.orafaq.net--Author: Freeman Robert - ILINET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L(or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).-- Please see the official ORACLE-L FAQ: http://www.orafaq.net-- Author: INET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Robert, you say yet like you are expecting problems? :) Tom Mercadante Oracle Certified Professional -Original Message- Sent: Wednesday, September 10, 2003 6:24 PM To: Multiple recipients of list ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Mercadante, Thomas F INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Just 2 minor ones on the upgrade so far - it had to have the compatible init parameter set - otherwise it tried to default to Ora 8 and I am using LMT's so it complained. Also I had a problem with the listener not being able to connect, so I deleted and recreated a new one - everything worked. Just now in the process of doing some regression testing with my test warehouse. So far nothing showing up. [EMAIL PROTECTED] 9/10/2003 5:24:28 PM Anyone have any negative experiences with 9.2.0.4 yet?Robert-- Please see the official ORACLE-L FAQ: http://www.orafaq.net-- Author: Freeman Robert - IL INET: [EMAIL PROTECTED]Fat City Network Services -- 858-538-5051 http://www.fatcity.comSan Diego, California -- Mailing list and web hosting services-To REMOVE yourself from this mailing list, send an E-Mail messageto: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and inthe message BODY, include a line containing: UNSUB ORACLE-L(or the name of mailing list you want to be removed from). You mayalso send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Linking issues with installing 9.2.0.4 patch on HP/UX... Had migrated a 8.1.7 database inplace to 9.2.0.2 whcih went fine and went to apply 9.2.0.4 patchset and the relink splattered all over the place. Patch requirments were vauge at best and found a list of required OS patches on HP's tech site but was unable to trackdown the source of those. Meta Link had some info but seemed unrelated to the specific linking errors I was getting. So now I am working with sysadmin in tracking down the patches. Needless to say I have a broken install waiting. This is our development box it has two other 8.1.7 instances and ERP running for developers. So there is some concern about the effect new OS patches will have on the old ERP software. Cannot stay at 9.2.0.2 due to HASH bug discussed here recently. Once I get the instance up we will be testing it pretty extensinvly and will be able to realte more on the database perf later. Have to get past the install first sounds fun eh.. -Original Message- Sent: Wednesday, September 10, 2003 5:24 PM To: Multiple recipients of list ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Odland, Brad INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Actually, I was wondering why Robert would care about 9.2.0.4 since his 10g New Features book (Oracle Press, 0-07-222947-0) is due out this December. Isn't 9i a little passé? :D Rich Rich Jesse System/Database Administrator [EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA -Original Message- From: Mercadante, Thomas F [mailto:[EMAIL PROTECTED] Sent: Thursday, September 11, 2003 7:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Robert, you say yet like you are expecting problems? :) Tom Mercadante Oracle Certified Professional -Original Message- Sent: Wednesday, September 10, 2003 6:24 PM To: Multiple recipients of list ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Jesse, Rich INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Hi! I wouldn't call it a real negative experience, but: C:\Work\Oraclesqlplus admin/admin SQL*Plus: Release 9.2.0.4.0 - Production on N Sep 11 17:21:59 2003 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.4.0 - Production With the Partitioning, OLAP and Oracle Data Mining options JServer Release 9.2.0.4.0 - Production SQL alter index i set; alter index i set * ERROR at line 1: ORA-03113: end-of-file on communication channel This statement writes some trace to core_dump_dest. The index name can be anything, seems that the syntax checking part crashes server process (or thread) since I'm on W2K SP4. Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Thursday, September 11, 2003 1:24 AM Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Tanel Poder INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
It's probably just someone executing some arbitrary code on your Windohs box. At least according to Microsoft Security Bulletin MS03-039 pushed out yesterday (Sep 10). Rich Rich Jesse System/Database Administrator [EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA -Original Message- From: Tanel Poder [mailto:[EMAIL PROTECTED] Sent: Thursday, September 11, 2003 10:30 AM To: Multiple recipients of list ORACLE-L Subject: Re: 9.2.0.4 anyone Hi! I wouldn't call it a real negative experience, but: C:\Work\Oraclesqlplus admin/admin SQL*Plus: Release 9.2.0.4.0 - Production on N Sep 11 17:21:59 2003 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.4.0 - Production With the Partitioning, OLAP and Oracle Data Mining options JServer Release 9.2.0.4.0 - Production SQL alter index i set; alter index i set * ERROR at line 1: ORA-03113: end-of-file on communication channel This statement writes some trace to core_dump_dest. The index name can be anything, seems that the syntax checking part crashes server process (or thread) since I'm on W2K SP4. Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Thursday, September 11, 2003 1:24 AM Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Jesse, Rich INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Oracle9.2.0.4 on Windows. I installed it yesterday and upgraded my test database, with OID, and it worked fine. It is available since 15AUG2003. Yechiel Adar Mehish - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Thursday, September 11, 2003 12:24 AM Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Yechiel Adar INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) _kghsidx_count = 1 that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
RE: 9.2.0.4 anyone
Title: RE: 9.2.0.4 anyone Parameter name _kghdsidx_count, default in 9.2.0.3 is 1, description max kghdsidx count. Do not know what it is for. Alex. -Original Message- From: Stephen Lee [mailto:[EMAIL PROTECTED]] Sent: Thursday, September 11, 2003 2:29 PM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) _kghsidx_count = 1 that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Hi! Oh yeah, I forgot to say that in 9.2.0.3 the shared pool was broken up into heaps (Oracle terminology) but whatever fancy stuff they were trying to accomplish by doing this (I think part of the magical self-tuning initiative) was buggy and this hastened the fatal 4031 situation. There is a parameter (so I have been told) Shared pool has always been a heap (inside SGA heap, and can contain heaps as well if requested). Shared pool heap is physically divided to equal sized extents. But space inside shared pool heap is allocated in chunks and there are freelistslru lists for tracking free and unpinned recreable tchunks. There is actually 255 freelists, each for different allocation size range, starting from 16 bytes up to about 64+ kilobytes. Depending on how much memory allocation request has been made, the proper freelist is scanned. When matching size chunks aren't found, the closest match is split. Anyway, lets call these 255 freelists a shared pool freelist set (don't know the correct Oracle term). But people often suffer from shared pool latch contention when a lot of shared pool memory allocation occurs (no bind variables are used in SQL etc etc). Lots of concurrent memory allocation requests mean lots of processes trying to acquire shared pool latch and scan through relevant freelist, if no sufficiently large chunk found, then LRU list as well. During these operations, the shared pool latch is being hold by session doing the scanning causing shared pool latch contention. Various experts suggesting to increase your shared pool make the situation even worse, because in time, the LRU and freelists get even longer, thus one scan takes more time to complete or fail. _kghsidx_count = 1 (The correct name is _kghdsidx_count) So, in 9i Oracle introduced the ability to create several heaps for serving shared pool. If you've set the shared_pool_size to 64M and _kghdsidx_count to 4, you'll have 4*16M shared pool heaps, every heap having its own descriptor, extents, LRU lists, freelists and shared pool child latch (there is no such child latch in 8i). Shorter lists mean faster scanning, more latches mean more scalability when serializing access to a resource. The downside is, that usually the resource has to be split that every latch protects one and only one part of resource. Thus if the shared pool is split into 4 parts and all memory allocation requests happen to use the same freelist for some reason, only 25% of memory can be used. This is the reason why you can avoid ORA-4031s when setting _kghsidx_count to 1 - this practically enables the old behaviour. In 8i there is only one set of freelists, in 9i there can be more sets, default is 1 and max limit is 7 as far as simple testing on my 9.2.0.4/W2k has showed - I have only 7 shared pool child latches and didn't find a way to increase them. When you increase _kghdsidx_count, you see more lines in x$kghlu as well, one for each heap (normally there was only one). Also you can verify the behaviour when taking SGA heapdump at level 2 and search for HEAP DUMP or FREE LISTS in trace, there are as many free lists in dump, as you've stated with _kghdsidx_count init parameter. (there is one extra heap in dump, this is the SGA parent heap for shared pool heaps). that makes the shared pool one big memory area. One of the fixes in 9.2.0.4 is to make this the default now (so I have been told). Can't get my hands oon 9.2.0.2 or 0.3, but in 0.4 (on Windows), the _kghsidx_count defaults to 1 anyway, check it out on your systems. Tanel. -Original Message- From: Stephen Lee Sent: Thursday, September 11, 2003 11:39 AM To: Multiple recipients of list ORACLE-L Subject: RE: 9.2.0.4 anyone Initial testing indicates that the bug(s) that caused index create/rebuild online to lock a table and then get permanently stuck in a hung state have been fixed. It looks like 9.2.0.4 does NOT fix the problem of fatal 4031 situations that can only be cleared by restarting the instance. So you are still stuck with by guess and by golly fiddling with the shared pool. -Original Message- Anyone have any negative experiences with 9.2.0.4 yet? -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Stephen Lee
RE: 9.2.0.4 anyone
I can get TOAD to hang almost consistently, but that may be the RAC part of 9.2.0.4 or my 9.2.0.1 client. I also (still) have a fervent dislike of the DBCA's mangling of given parameters -- even when it doesn't go off into la-la land (it hangs at Initializing... about 50% of the time). Other than that, I haven't seen any bugs in the DB itself -- yet. I've only been testing data-wise this week. Rich Rich Jesse System/Database Administrator [EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA -Original Message- From: Freeman Robert - IL [mailto:[EMAIL PROTECTED] Sent: Wednesday, September 10, 2003 5:24 PM To: Multiple recipients of list ORACLE-L Subject: 9.2.0.4 anyone Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Jesse, Rich INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Now that you speak about TOAD I remember that I get ORA-01460: unimplemented or unreasonable conversion requested in Toads schema browser. But I'm not using it much anyway, thus don't care... Tanel. - Original Message - To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Sent: Thursday, September 11, 2003 1:54 AM I can get TOAD to hang almost consistently, but that may be the RAC part of 9.2.0.4 or my 9.2.0.1 client. I also (still) have a fervent dislike of the DBCA's mangling of given parameters -- even when it doesn't go off into la-la land (it hangs at Initializing... about 50% of the time). Other than that, I haven't seen any bugs in the DB itself -- yet. I've only been testing data-wise this week. Rich Rich Jesse System/Database Administrator [EMAIL PROTECTED] Quad/Tech Inc, Sussex, WI USA -Original Message- From: Freeman Robert - IL [mailto:[EMAIL PROTECTED] Sent: Wednesday, September 10, 2003 5:24 PM To: Multiple recipients of list ORACLE-L Subject: 9.2.0.4 anyone Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Jesse, Rich INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Tanel Poder INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
What? 9.2.0.4 is out now? RWB Reginald W. Bailey IBM Global Services - ETS SW GDSD - Database Management Your Friendly Neighborhood DBA 713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager) [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED] Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: INET: [EMAIL PROTECTED] Fat City Network Services-- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: 9.2.0.4 anyone
Yeah, been out for some time now actually. [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 09/10/2003 04:25 PM Please respond to ORACLE-L To:Multiple recipients of list ORACLE-L [EMAIL PROTECTED] cc: Subject:Re: 9.2.0.4 anyone What? 9.2.0.4 is out now? RWB Reginald W. Bailey IBM Global Services - ETS SW GDSD - Database Management Your Friendly Neighborhood DBA 713-216-7703 (Office) 281-798-5474 (Mobile) 713-415-5410 (Pager) [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] com To: [EMAIL PROTECTED] Sent by: cc: [EMAIL PROTECTED]Subject: 9.2.0.4 anyone ity.com 09/10/2003 05:24 PM Please respond to ORACLE-L Anyone have any negative experiences with 9.2.0.4 yet? Robert -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: Freeman Robert - IL INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.net -- Author: INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California-- Mailing list and web hosting services - To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
