tempdbs used by a ghost
Alberto (IDS 11.50.FC7W1GE on RHEL x86_64) saw two of three temp dbspaces filling up with no identifiable session. Alexandre supplied a sysmaster query (systabnames/systabinfo with ti_flags bit tests) that revealed thousands of orphaned SORTTEMP th_tmprun tables owned by a killed user session; the online.log showed an "Assert Failed: pool not freed" after that session died. Art explained orphaned sort space normally frees when the sort finishes, and suggested rebooting the user's PC to release a hung client connection. Dave identified it as bug IC73918, fixed in 11.50.xC9, noting the space can only be reclaimed by bouncing the engine, with extra temp chunks as a stopgap. No confirmation of the outcome is recorded.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi folks! I Couldn't identify a session who is using the temporary DBSpace of my company's system. The SNMP Monitoring Tool is pointing that the free space of 2 of 3 spaces has been used, and I can't find who is using. PS: I can't bounce the engine this week to solve this problem... Can anybody help me? Thanks in advance, Alberto.
Hi Alberto. Did you try running this sysmaster query? SELECT trim(n.dbsname), trim(n.owner), trim(n.tabname), COUNT(*), SUM(i.ti_nptotal) as pages, SUM(i.ti_nrows) as rows FROM systabnames n, systabinfo i WHERE (BITVAL(i.ti_flags, "0x0020") = 1 OR BITVAL(i.ti_flags, "0x0040") = 1) AND i.ti_partnum = n.partnum group by 1,2,3 --ORDER BY 5 DESC You didn´t mention your engine and platform, so I hope it works. It works fine on my linux_64 with 11.50.FC8W3 engine. Regards. Em 14/09/2011 09:22, ALBERTO ROMEU PESSONIO FILHO escreveu: > Hi folks! > > I Couldn't identify a session who is using the temporary DBSpace of my > company's system. The SNMP Monitoring Tool is pointing that the free space of > 2 of 3 spaces has been used, and I can't find who is using. > > PS: I can't bounce the engine this week to solve this problem... > > Can anybody help me? > Thanks in advance, > Alberto. > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Alexandre Marini Tecnologia da Informação - DBA SEFAZ-MS / SGI-UGSR / Sistemas IBM-Informix <Cert-Info-Mgmt_color.jpg> IBM Certified System Administrator - Informix Dynamic Server V10 / V11 / V11.70 IBM Information Management Informix Technical Professional v3
It might be a running archive process. The archive copies physical log records for each dbspace into a temp table that is fragmented across the temp dbspaces. What version of Informix are you using and on what platform? That may shed some light. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 9:22 AM, ALBERTO ROMEU PESSONIO FILHO < arfilho@orizonbrasil.com.br> wrote: > Hi folks! > > I Couldn't identify a session who is using the temporary DBSpace of my > company's system. The SNMP Monitoring Tool is pointing that the free space > of > 2 of 3 spaces has been used, and I can't find who is using. > > PS: I can't bounce the engine this week to solve this problem... > > Can anybody help me? > Thanks in advance, > Alberto. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --20cf303ea5ce22409704ace6d721
Hi guys!!! Sorry to forget this: IDS 11.50 FC7W1GE Linux Redhat x86_64bits Art: My system wasn't in archive mode when the issue happened. I suspect that we are running almost out of memory on the Growth Edition License, and another session tried to get memory from the virtual segment. Look at my Online.log: 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) user:rrosini, prog:, ses id:44429442 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, address:0x0x2edf86040 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) Thread(288624763, sqlexec, 3138a02e8, 1) File: session.c Line: 2062 09/13/11 14:43:08 Results: Internally corrected 09/13/11 14:43:08 stack trace for pid 17741 written to /informix/logs_informix/ifxorizon1/dump_af/af.1463962c 09/13/11 14:43:08 See Also: /informix/logs_informix/ifxorizon1/dump_af/af.1463962c 09/13/11 14:43:45 Releasing server from system block 09/13/11 14:43:45 Releasing server from system block 09/13/11 14:43:45 Releasing server from system block 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This can be due to slow network file access. Will try 12 more times 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, address:0x0x2edf86040 09/13/11 14:49:43 VP Notify mechanism completed This session (rrosini) was running a query from Aqua Data Studio Tool near at noon, and 14:43 this error happened. this incident caused five minutes of system instability, but I had no action. The problem is: I've killed the session, but the engine didn't freed the temporary DBSpace, and I can't attach the space used on the temp DBSpace to the user rrosini The query that Alexandre sent me helped a lot. Thank you very much! I executed the query using a filter for 'rrosini'. The output brought me 12.922 lines. SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 How can I freed this space to the engine without bounce it? Thanks, Alberto -----Mensagem original----- De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art Kagel Enviada em: quarta-feira, 14 de setembro de 2011 10:38 Para: ids@iiug.org Assunto: Re: tempdbs used by a ghost [24905] It might be a running archive process. The archive copies physical log records for each dbspace into a temp table that is fragmented across the temp dbspaces. What version of Informix are you using and on what platform? That may shed some light. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 9:22 AM, ALBERTO ROMEU PESSONIO FILHO < arfilho@orizonbrasil.com.br> wrote: > Hi folks! > > I Couldn't identify a session who is using the temporary DBSpace of my > company's system. The SNMP Monitoring Tool is pointing that the free space > of > 2 of 3 spaces has been used, and I can't find who is using. > > PS: I can't bounce the engine this week to solve this problem... > > Can anybody help me? > Thanks in advance, > Alberto. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --20cf303ea5ce22409704ace6d721 ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
If the rrosini session is already gone then there isn't anything you can do to get rid of the orphaned temp space short of bouncing the instance. Usually it is still lying around because the sort is still going on (could be a pathological query like a cartesian product sorting trillions of rows) and the space will free once the sort completes and the engine tries to tell the session that data is finally available to fetch. Either that or it will run out of temp dbspace and that will cause the query to rollback. Either way at that point it will realize that the session is gone and release the resources the session is holding. However, if the query that is sorting it VERY large that may not happen for quite some time. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi guys!!! > > Sorry to forget this: > > IDS 11.50 FC7W1GE > Linux Redhat x86_64bits > > Art: My system wasn't in archive mode when the issue happened. > > I suspect that we are running almost out of memory on the Growth Edition > License, and another session tried to get memory from the virtual segment. > > Look at my Online.log: > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > user:rrosini, prog:, ses id:44429442 > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > address:0x0x2edf86040 > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > Thread(288624763, sqlexec, 3138a02e8, 1) > > File: session.c Line: 2062 > 09/13/11 14:43:08 Results: Internally corrected > 09/13/11 14:43:08 stack trace for pid 17741 written to > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > 09/13/11 14:43:08 See Also: > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This can > be > due to slow network file access. Will try 12 more times > 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, > address:0x0x2edf86040 > 09/13/11 14:49:43 VP Notify mechanism completed > > This session (rrosini) was running a query from Aqua Data Studio Tool near > at > noon, and 14:43 this error happened. this incident caused five minutes of > system instability, but I had no action. > > The problem is: I've killed the session, but the engine didn't freed the > temporary DBSpace, and I can't attach the space used on the temp DBSpace to > the user rrosini > > The query that Alexandre sent me helped a lot. Thank you very much! > I executed the query using a filter for 'rrosini'. The output brought me > 12.922 lines. > > SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 > SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 > SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 > SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 > SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 > SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 > SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 > SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 > > How can I freed this space to the engine without bounce it? > > Thanks, > Alberto > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 10:38 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24905] > > It might be a running archive process. The archive copies physical log > records for each dbspace into a temp table that is fragmented across the > temp dbspaces. What version of Informix are you using and on what > platform? That may shed some light. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 9:22 AM, ALBERTO ROMEU PESSONIO FILHO < > arfilho@orizonbrasil.com.br> wrote: > > > Hi folks! > > > > I Couldn't identify a session who is using the temporary DBSpace of my > > company's system. The SNMP Monitoring Tool is pointing that the free > space > > of > > 2 of 3 spaces has been used, and I can't find who is using. > > > > PS: I can't bounce the engine this week to solve this problem... > > > > Can anybody help me? > > Thanks in advance, > > Alberto. > > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > --20cf303ea5ce22409704ace6d721 > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --20cf303ea5ce60427604ace75f69
Hi Art! I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she got the result yesterday. I think the query fetched data and finished. Also, I saw the query, and this kind of query when is triggered on the instance usually brings rows in 3 minutes, at max. I think that I really need to bounce the engine. Thanks in advance, Alberto. -----Mensagem original----- De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art Kagel Enviada em: quarta-feira, 14 de setembro de 2011 11:14 Para: ids@iiug.org Assunto: Re: tempdbs used by a ghost [24908] If the rrosini session is already gone then there isn't anything you can do to get rid of the orphaned temp space short of bouncing the instance. Usually it is still lying around because the sort is still going on (could be a pathological query like a cartesian product sorting trillions of rows) and the space will free once the sort completes and the engine tries to tell the session that data is finally available to fetch. Either that or it will run out of temp dbspace and that will cause the query to rollback. Either way at that point it will realize that the session is gone and release the resources the session is holding. However, if the query that is sorting it VERY large that may not happen for quite some time. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi guys!!! > > Sorry to forget this: > > IDS 11.50 FC7W1GE > Linux Redhat x86_64bits > > Art: My system wasn't in archive mode when the issue happened. > > I suspect that we are running almost out of memory on the Growth Edition > License, and another session tried to get memory from the virtual segment. > > Look at my Online.log: > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > user:rrosini, prog:, ses id:44429442 > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > address:0x0x2edf86040 > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > Thread(288624763, sqlexec, 3138a02e8, 1) > > File: session.c Line: 2062 > 09/13/11 14:43:08 Results: Internally corrected > 09/13/11 14:43:08 stack trace for pid 17741 written to > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > 09/13/11 14:43:08 See Also: > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:43:45 Releasing server from system block > 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This can > be > due to slow network file access. Will try 12 more times > 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, > address:0x0x2edf86040 > 09/13/11 14:49:43 VP Notify mechanism completed > > This session (rrosini) was running a query from Aqua Data Studio Tool near > at > noon, and 14:43 this error happened. this incident caused five minutes of > system instability, but I had no action. > > The problem is: I've killed the session, but the engine didn't freed the > temporary DBSpace, and I can't attach the space used on the temp DBSpace to > the user rrosini > > The query that Alexandre sent me helped a lot. Thank you very much! > I executed the query using a filter for 'rrosini'. The output brought me > 12.922 lines. > > SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 > SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 > SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 > SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 > SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 > SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 > SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 > SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 > SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 > SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 > SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 > > How can I freed this space to the engine without bounce it? > > Thanks, > Alberto > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 10:38 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24905] > > It might be a running archive process. The archive copies physical log > records for each dbspace into a temp table that is fragmented across the > temp dbspaces. What version of Informix are you using and on what > platform? That may shed some light. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 9:22 AM, ALBERTO ROMEU PESSONIO FILHO < > arfilho@orizonbrasil.com.br> wrote: > > > Hi folks! > > > > I Couldn't identify a session who is using the temporary DBSpace of my > > company's system. The SNMP Monitoring Tool is pointing that the free > space > > of > > 2 of 3 spaces has been used, and I can't find who is using. > > > > PS: I can't bounce the engine this week to solve this problem... > > > > Can anybody help me? > > Thanks in advance, > > Alberto. > > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > --20cf303ea5ce22409704ace6d721 > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > ******************************************************************************* > Forum Note: Use "Reply@@
Alberto Romeu Pessonio Filho Wrote: -------------------------------------------------------------------------------- Hi guys!!! Sorry to forget this: IDS 11.50 FC7W1GE Linux Redhat x86_64bits Art: My system wasn't in archive mode when the issue happened. I suspect that we are running almost out of memory on the Growth Edition License, and another session tried to get memory from the virtual segment. Look at my Online.log: 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) user:rrosini, prog:, ses id:44429442 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, address:0x0x2edf86040 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) Thread(288624763, sqlexec, 3138a02e8, 1) File: session.c Line: 2062 ...snip... How can I freed this space to the engine without bounce it? Thanks, Alberto -------------------------------------------------------------------------------- Response: Might be this bug... IC73918 ASSERTION IN FREE_POOLS() WHEN MORE THAN ONE SORT FILE EXISTS AND THE PROCESS WAS STOPPED OR FILE SYSTEM BECAME FULL It has been addressed in 11.50.xC9. You need to bounce the engine to clear the space. If the business impact of a bounce is too great, you could add chunks to your tempspaces to delay the need. Dave Griffen
One thing to try first: If rrosini ran the query from a PC based tool, try rebooting her PC first. That may release the network connection. PC apps are terrible about properly shutting down their network connections and sometimes that leaves Informix blocked waiting for a response from an app that is long gone. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi Art! > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she got > the result yesterday. I think the query fetched data and finished. Also, I > saw > the query, and this kind of query when is triggered on the instance usually > brings rows in 3 minutes, at max. > > I think that I really need to bounce the engine. > > Thanks in advance, > Alberto. > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 11:14 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24908] > > If the rrosini session is already gone then there isn't anything you can do > to get rid of the orphaned temp space short of bouncing the instance. > Usually it is still lying around because the sort is still going on (could > be a pathological query like a cartesian product sorting trillions of rows) > and the space will free once the sort completes and the engine tries to > tell > the session that data is finally available to fetch. Either that or it will > run out of temp dbspace and that will cause the query to rollback. Either > way at that point it will realize that the session is gone and release the > resources the session is holding. However, if the query that is sorting it > VERY large that may not happen for quite some time. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < > arfilho@orizonbrasil.com.br> wrote: > > > Hi guys!!! > > > > Sorry to forget this: > > > > IDS 11.50 FC7W1GE > > Linux Redhat x86_64bits > > > > Art: My system wasn't in archive mode when the issue happened. > > > > I suspect that we are running almost out of memory on the Growth Edition > > License, and another session tried to get memory from the virtual > segment. > > > > Look at my Online.log: > > > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > > user:rrosini, prog:, ses id:44429442 > > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > > address:0x0x2edf86040 > > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > > > Thread(288624763, sqlexec, 3138a02e8, 1) > > > > File: session.c Line: 2062 > > 09/13/11 14:43:08 Results: Internally corrected > > 09/13/11 14:43:08 stack trace for pid 17741 written to > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > 09/13/11 14:43:08 See Also: > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This > can > > be > > due to slow network file access. Will try 12 more times > > 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, > > address:0x0x2edf86040 > > 09/13/11 14:49:43 VP Notify mechanism completed > > > > This session (rrosini) was running a query from Aqua Data Studio Tool > near > > at > > noon, and 14:43 this error happened. this incident caused five minutes of > > system instability, but I had no action. > > > > The problem is: I've killed the session, but the engine didn't freed the > > temporary DBSpace, and I can't attach the space used on the temp DBSpace > to > > the user rrosini > > > > The query that Alexandre sent me helped a lot. Thank you very much! > > I executed the query using a filter for 'rrosini'. The output brought me > > 12.922 lines. > > > > SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 > > SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 > > > > How can I freed this space to the engine without bounce it? > > > > Thanks, > > Alberto > > > > -----Mensagem original----- > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > > Kagel > > Enviada em: quarta-feira, 14 de setembro de 2011 10:38 > > Para: ids@iiug.org > > Assunto: Re: tempdbs used by a ghost [24905] > > > > It might be a running archive process. The archive copies physical log > > records for each dbspace into a temp table that is fragmented across the > > temp dbspaces. What version of Informix are you using and on what > > platform? That may shed some light. > > > > Art > > > > Art S. Kagel > > Advanced DataTools (www.advancedatatools.com) > > Blog: http://informix-myview.blogspot.com/ > > > > Disclaimer: Please keep in mind that my own opinions are my own opinions > > and > > do not reflect on my employer, Advanced DataTools, the IIUG, nor any > other > > organization with which I am associated either explicitly, implicitly, or > > by > > inference. Ne
Yes, that´s a nice try. I had several issues here, coming from java clients doing XA transactions that never freeds table locks. It´s a little difficult to find, but nothing that a good "technical connection configuration manual" doesn´t solve in a glance... kkkkkk Em 14/09/2011 12:06, Art Kagel escreveu: > One thing to try first: If rrosini ran the query from a PC based tool, try > rebooting her PC first. That may release the network connection. PC apps > are terrible about properly shutting down their network connections and > sometimes that leaves Informix blocked waiting for a response from an app > that is long gone. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho< > arfilho@orizonbrasil.com.br> wrote: > >> Hi Art! >> >> I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she got >> the result yesterday. I think the query fetched data and finished. Also, I >> saw >> the query, and this kind of query when is triggered on the instance usually >> brings rows in 3 minutes, at max. >> >> I think that I really need to bounce the engine. >> >> Thanks in advance, >> Alberto. >> >> -----Mensagem original----- >> De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art >> Kagel >> Enviada em: quarta-feira, 14 de setembro de 2011 11:14 >> Para: ids@iiug.org >> Assunto: Re: tempdbs used by a ghost [24908] >> >> If the rrosini session is already gone then there isn't anything you can do >> to get rid of the orphaned temp space short of bouncing the instance. >> Usually it is still lying around because the sort is still going on (could >> be a pathological query like a cartesian product sorting trillions of rows) >> and the space will free once the sort completes and the engine tries to >> tell >> the session that data is finally available to fetch. Either that or it will >> run out of temp dbspace and that will cause the query to rollback. Either >> way at that point it will realize that the session is gone and release the >> resources the session is holding. However, if the query that is sorting it >> VERY large that may not happen for quite some time. >> >> Art >> >> Art S. Kagel >> Advanced DataTools (www.advancedatatools.com) >> Blog: http://informix-myview.blogspot.com/ >> >> Disclaimer: Please keep in mind that my own opinions are my own opinions >> and >> do not reflect on my employer, Advanced DataTools, the IIUG, nor any other >> organization with which I am associated either explicitly, implicitly, or >> by >> inference. Neither do those opinions reflect those of other individuals >> affiliated with any entity with which I am affiliated nor those of the >> entities themselves. >> >> On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho< >> arfilho@orizonbrasil.com.br> wrote: >> >>> Hi guys!!! >>> >>> Sorry to forget this: >>> >>> IDS 11.50 FC7W1GE >>> Linux Redhat x86_64bits >>> >>> Art: My system wasn't in archive mode when the issue happened. >>> >>> I suspect that we are running almost out of memory on the Growth Edition >>> License, and another session tried to get memory from the virtual >> segment. >>> Look at my Online.log: >>> >>> 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) >>> user:rrosini, prog:, ses id:44429442 >>> 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, >>> address:0x0x2edf86040 >>> 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE >>> 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) >>> >>> Thread(288624763, sqlexec, 3138a02e8, 1) >>> >>> File: session.c Line: 2062 >>> 09/13/11 14:43:08 Results: Internally corrected >>> 09/13/11 14:43:08 stack trace for pid 17741 written to >>> /informix/logs_informix/ifxorizon1/dump_af/af.1463962c >>> 09/13/11 14:43:08 See Also: >>> /informix/logs_informix/ifxorizon1/dump_af/af.1463962c >>> 09/13/11 14:43:45 Releasing server from system block >>> 09/13/11 14:43:45 Releasing server from system block >>> 09/13/11 14:43:45 Releasing server from system block >>> 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This >> can >>> be >>> due to slow network file access. Will try 12 more times >>> 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, >>> address:0x0x2edf86040 >>> 09/13/11 14:49:43 VP Notify mechanism completed >>> >>> This session (rrosini) was running a query from Aqua Data Studio Tool >> near >>> at >>> noon, and 14:43 this error happened. this incident caused five minutes of >>> system instability, but I had no action. >>> >>> The problem is: I've killed the session, but the engine didn't freed the >>> temporary DBSpace, and I can't attach the space used on the temp DBSpace >> to >>> the user rrosini >>> >>> The query that Alexandre sent me helped a lot. Thank you very much! >>> I executed the query using a filter for 'rrosini'. The output brought me >>> 12.922 lines. >>> >>> SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 >>> SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 >>> >>> How can I freed this space to the engine without bounce it? >>> >>> Thanks, >>> Alberto >>> >>> -----Mensagem original----- >>> De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art >>> Kagel >>> Enviada em: quarta-feira, 14 de setembro de 2011 10:38 >>> Para: ids@iiug.org >>> Assunto: Re: tempdbs used by a ghost [24905] >>> >>> It might be a running archive process. The archive copies physical log >>> records for each dbspace into a temp table that is fragmented across the >>> temp dbspaces. What version of Informix are you using and on what >>> platform? That may shed some light. >>> >>> Art >>> >
Hi Art... I tried restart the PC of rrosini, but it not solved. Unfortunately, i'll restart the engine late at night and apply 11.50 FC9 release, as Dave suggested. The situation is identical as the bug reports. Thank you guys!!! Best regards, Alberto. -----Mensagem original----- De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art Kagel Enviada em: quarta-feira, 14 de setembro de 2011 13:06 Para: ids@iiug.org Assunto: Re: tempdbs used by a ghost [24914] One thing to try first: If rrosini ran the query from a PC based tool, try rebooting her PC first. That may release the network connection. PC apps are terrible about properly shutting down their network connections and sometimes that leaves Informix blocked waiting for a response from an app that is long gone. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi Art! > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she got > the result yesterday. I think the query fetched data and finished. Also, I > saw > the query, and this kind of query when is triggered on the instance usually > brings rows in 3 minutes, at max. > > I think that I really need to bounce the engine. > > Thanks in advance, > Alberto. > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 11:14 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24908] > > If the rrosini session is already gone then there isn't anything you can do > to get rid of the orphaned temp space short of bouncing the instance. > Usually it is still lying around because the sort is still going on (could > be a pathological query like a cartesian product sorting trillions of rows) > and the space will free once the sort completes and the engine tries to > tell > the session that data is finally available to fetch. Either that or it will > run out of temp dbspace and that will cause the query to rollback. Either > way at that point it will realize that the session is gone and release the > resources the session is holding. However, if the query that is sorting it > VERY large that may not happen for quite some time. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < > arfilho@orizonbrasil.com.br> wrote: > > > Hi guys!!! > > > > Sorry to forget this: > > > > IDS 11.50 FC7W1GE > > Linux Redhat x86_64bits > > > > Art: My system wasn't in archive mode when the issue happened. > > > > I suspect that we are running almost out of memory on the Growth Edition > > License, and another session tried to get memory from the virtual > segment. > > > > Look at my Online.log: > > > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > > user:rrosini, prog:, ses id:44429442 > > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > > address:0x0x2edf86040 > > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > > > Thread(288624763, sqlexec, 3138a02e8, 1) > > > > File: session.c Line: 2062 > > 09/13/11 14:43:08 Results: Internally corrected > > 09/13/11 14:43:08 stack trace for pid 17741 written to > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > 09/13/11 14:43:08 See Also: > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:43:45 Releasing server from system block > > 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This > can > > be > > due to slow network file access. Will try 12 more times > > 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, > > address:0x0x2edf86040 > > 09/13/11 14:49:43 VP Notify mechanism completed > > > > This session (rrosini) was running a query from Aqua Data Studio Tool > near > > at > > noon, and 14:43 this error happened. this incident caused five minutes of > > system instability, but I had no action. > > > > The problem is: I've killed the session, but the engine didn't freed the > > temporary DBSpace, and I can't attach the space used on the temp DBSpace > to > > the user rrosini > > > > The query that Alexandre sent me helped a lot. Thank you very much! > > I executed the query using a filter for 'rrosini'. The output brought me > > 12.922 lines. > > > > SORTTEMP rrosini th_tmprun_0x2fb99f578 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2fb99f8e8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2fcc7bde8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5078 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5348 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5618 1 576 0 > > SORTTEMP rrosini th_tmprun_0x31bda5fc8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2ee8cc578 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2ee8cc7a8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x386e8f9d8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2efa6c078 1 576 0 > > SORTTEMP rrosini th_tmprun_0x302ee77f8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3ba928938 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3ba928ed8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x3030f0398 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860d438 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860db68 1 576 0 > > SORTTEMP rrosini th_tmprun_0x30860dfc8 1 576 0 > > SORTTEMP rrosini th_tmprun_0x2f815a1b8 1 576 0 > > > > How can I freed this space to the engine without bounce it? > > > > Thanks, > > Alberto > > > > -----Mensagem original----- > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > > Kagel > > Enviada em: quarta-feira, 14 de setembro de 2011 10:38 > > Para: ids@iiug.org > > Assunto: Re: tempdbs used by a ghost [24905] > > > > It might be a running archive process. The archive copies physical log > > records for each dbspace into a temp table that is fragmented across the > > temp dbspaces. Wh
Best practice: - Get 11.50.FC9 installed in a new directory during the day and copy in your ONCONFIG and sqlhosts files ready to go, and link any subdirectories that should be under the new INFORMIXDIR after the restart into the .FC9 installation. - Then when you have to bounce your downtime will just be a minute or three to go down, relink INFORMIXDIR to point to the FC9 installation, and restart 11.50.FC9. Also if there is some problem starting .FC9, FC8 is still ready to go back online. - Go to sleep happy and comfortable. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 1:58 PM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi Art... > > I tried restart the PC of rrosini, but it not solved. > Unfortunately, i'll restart the engine late at night and apply 11.50 FC9 > release, as Dave suggested. The situation is identical as the bug reports. > > Thank you guys!!! > > Best regards, > Alberto. > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 13:06 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24914] > > One thing to try first: If rrosini ran the query from a PC based tool, try > rebooting her PC first. That may release the network connection. PC apps > are terrible about properly shutting down their network connections and > sometimes that leaves Informix blocked waiting for a response from an app > that is long gone. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho < > arfilho@orizonbrasil.com.br> wrote: > > > Hi Art! > > > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she > got > > the result yesterday. I think the query fetched data and finished. Also, > I > > saw > > the query, and this kind of query when is triggered on the instance > usually > > brings rows in 3 minutes, at max. > > > > I think that I really need to bounce the engine. > > > > Thanks in advance, > > Alberto. > > > > -----Mensagem original----- > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > > Kagel > > Enviada em: quarta-feira, 14 de setembro de 2011 11:14 > > Para: ids@iiug.org > > Assunto: Re: tempdbs used by a ghost [24908] > > > > If the rrosini session is already gone then there isn't anything you can > do > > to get rid of the orphaned temp space short of bouncing the instance. > > Usually it is still lying around because the sort is still going on > (could > > be a pathological query like a cartesian product sorting trillions of > rows) > > and the space will free once the sort completes and the engine tries to > > tell > > the session that data is finally available to fetch. Either that or it > will > > run out of temp dbspace and that will cause the query to rollback. Either > > way at that point it will realize that the session is gone and release > the > > resources the session is holding. However, if the query that is sorting > it > > VERY large that may not happen for quite some time. > > > > Art > > > > Art S. Kagel > > Advanced DataTools (www.advancedatatools.com) > > Blog: http://informix-myview.blogspot.com/ > > > > Disclaimer: Please keep in mind that my own opinions are my own opinions > > and > > do not reflect on my employer, Advanced DataTools, the IIUG, nor any > other > > organization with which I am associated either explicitly, implicitly, or > > by > > inference. Neither do those opinions reflect those of other individuals > > affiliated with any entity with which I am affiliated nor those of the > > entities themselves. > > > > On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < > > arfilho@orizonbrasil.com.br> wrote: > > > > > Hi guys!!! > > > > > > Sorry to forget this: > > > > > > IDS 11.50 FC7W1GE > > > Linux Redhat x86_64bits > > > > > > Art: My system wasn't in archive mode when the issue happened. > > > > > > I suspect that we are running almost out of memory on the Growth > Edition > > > License, and another session tried to get memory from the virtual > > segment. > > > > > > Look at my Online.log: > > > > > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > > > user:rrosini, prog:, ses id:44429442 > > > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > > > address:0x0x2edf86040 > > > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > > > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > > > > > Thread(288624763, sqlexec, 3138a02e8, 1) > > > > > > File: session.c Line: 2062 > > > 09/13/11 14:43:08 Results: Internally corrected > > > 09/13/11 14:43:08 stack trace for pid 17741 written to > > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > > 09/13/11 14:43:08 See Also: > > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > > 09/13/11 14:43:45 Releasing server from system block > > > 09/13/11 14:43:45 Releasing server from system block > > > 09/13/11 14:43:45 Releasing server from system block > > > 09/13/11 14:48:44 VP Notify mechanism incomplete after 5 minutes. This > > can > > > be > > > due to slow network file access. Will try 12 more times > > > 09/13/11 14:49:43 pool not freed. pool name:44429442_SO, > > > address:0x0x2edf86040 > > > 09/13/11 14:49:43 VP Notify mechanism completed > > > > > > This session (rrosini) was running a query from Aqua Data Studio Tool > > near > > > at > > > noon, and 14:43 this error happened. this incident caused five minutes > of > > > system instability, but I had no action. > > > > > > The problem is: I've killed the session, but the engine didn't freed > the > > > temporary DBSpace, and I can't attach the space used on the temp > DBSpace > > to > > > the user rrosini > > > > > > The query that Alexandre sent me helped a lot. Thank you very much! > > > I executed the query using a filter for 'rrosini'. Th
OK Art!!! But, in previous minor versions of IDS 11.50, when we apply new commercial versions (i.e: from 11.5 FC5 to 11.5 FC6) I noticed we had Page migrations. If we have page migrations, something is gonna change on a header structure of the Pages. In this case, we might cause decrease performance IF the maintenance hasn't a good analysis and plan. Although this change is easy to do, I need to have some precaution because it's a 24X7 core system business with a little time response to the application (billing authorization system). If something goes wrong, my head is gonna be hunted at my company. Anyway, I thank your suggestion, but it will be done in couple weeks. Thank you very much, Alberto. -----Mensagem original----- De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art Kagel Enviada em: quarta-feira, 14 de setembro de 2011 15:18 Para: ids@iiug.org Assunto: Re: tempdbs used by a ghost [24925] Best practice: - Get 11.50.FC9 installed in a new directory during the day and copy in your ONCONFIG and sqlhosts files ready to go, and link any subdirectories that should be under the new INFORMIXDIR after the restart into the .FC9 installation. - Then when you have to bounce your downtime will just be a minute or three to go down, relink INFORMIXDIR to point to the FC9 installation, and restart 11.50.FC9. Also if there is some problem starting .FC9, FC8 is still ready to go back online. - Go to sleep happy and comfortable. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Wed, Sep 14, 2011 at 1:58 PM, Alberto Romeu Pessonio Filho < arfilho@orizonbrasil.com.br> wrote: > Hi Art... > > I tried restart the PC of rrosini, but it not solved. > Unfortunately, i'll restart the engine late at night and apply 11.50 FC9 > release, as Dave suggested. The situation is identical as the bug reports. > > Thank you guys!!! > > Best regards, > Alberto. > > -----Mensagem original----- > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > Kagel > Enviada em: quarta-feira, 14 de setembro de 2011 13:06 > Para: ids@iiug.org > Assunto: Re: tempdbs used by a ghost [24914] > > One thing to try first: If rrosini ran the query from a PC based tool, try > rebooting her PC first. That may release the network connection. PC apps > are terrible about properly shutting down their network connections and > sometimes that leaves Informix blocked waiting for a response from an app > that is long gone. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and > do not reflect on my employer, Advanced DataTools, the IIUG, nor any other > organization with which I am associated either explicitly, implicitly, or > by > inference. Neither do those opinions reflect those of other individuals > affiliated with any entity with which I am affiliated nor those of the > entities themselves. > > On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho < > arfilho@orizonbrasil.com.br> wrote: > > > Hi Art! > > > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she > got > > the result yesterday. I think the query fetched data and finished. Also, > I > > saw > > the query, and this kind of query when is triggered on the instance > usually > > brings rows in 3 minutes, at max. > > > > I think that I really need to bounce the engine. > > > > Thanks in advance, > > Alberto. > > > > -----Mensagem original----- > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art > > Kagel > > Enviada em: quarta-feira, 14 de setembro de 2011 11:14 > > Para: ids@iiug.org > > Assunto: Re: tempdbs used by a ghost [24908] > > > > If the rrosini session is already gone then there isn't anything you can > do > > to get rid of the orphaned temp space short of bouncing the instance. > > Usually it is still lying around because the sort is still going on > (could > > be a pathological query like a cartesian product sorting trillions of > rows) > > and the space will free once the sort completes and the engine tries to > > tell > > the session that data is finally available to fetch. Either that or it > will > > run out of temp dbspace and that will cause the query to rollback. Either > > way at that point it will realize that the session is gone and release > the > > resources the session is holding. However, if the query that is sorting > it > > VERY large that may not happen for quite some time. > > > > Art > > > > Art S. Kagel > > Advanced DataTools (www.advancedatatools.com) > > Blog: http://informix-myview.blogspot.com/ > > > > Disclaimer: Please keep in mind that my own opinions are my own opinions > > and > > do not reflect on my employer, Advanced DataTools, the IIUG, nor any > other > > organization with which I am associated either explicitly, implicitly, or > > by > > inference. Neither do those opinions reflect those of other individuals > > affiliated with any entity with which I am affiliated nor those of the > > entities themselves. > > > > On Wed, Sep 14, 2011 at 9:56 AM, Alberto Romeu Pessonio Filho < > > arfilho@orizonbrasil.com.br> wrote: > > > > > Hi guys!!! > > > > > > Sorry to forget this: > > > > > > IDS 11.50 FC7W1GE > > > Linux Redhat x86_64bits > > > > > > Art: My system wasn't in archive mode when the issue happened. > > > > > > I suspect that we are running almost out of memory on the Growth > Edition > > > License, and another session tried to get memory from the virtual > > segment. > > > > > > Look at my Online.log: > > > > > > 09/13/11 14:43:08 WARNING: 1 mem pools not freed. (destroy_session) > > > user:rrosini, prog:, ses id:44429442 > > > 09/13/11 14:43:08 Assert Failed: pool not freed. pool name:44429442_SO, > > > address:0x0x2edf86040 > > > 09/13/11 14:43:08 IBM Informix Dynamic Server Version 11.50.FC7W1GE > > > 09/13/11 14:43:08 Who: Session(44429442, rrosini@æx, -1, 0x2e8366a50) > > > > > > Thread(288624763, sqlexec, 3138a02e8, 1) > > > > > > File: session.c Line: 2062 > > > 09/13/11 14:43:08 Results: Internally corrected > > > 09/13/11 14:43:08 stack trace for pid 17741 written to > > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > > 09/13/11 14:43:08 See Also: > > > /informix/logs_informix/ifxorizon1/dump_af/af.1463962c > > > 09/13/11 14:43:45 Releasing server from system block > > > 09/13/11 14:43:45 Releasing server from system block >
Understandable and understood. The new features and structural changes in
11.50 stopped with 11.50.xC6 or 7 when 11.70 took over, so you should be
OK. That said, good DBAs are paranoid DBAs, so caution is the watchword and
planning too. It's always a good idea to run drop all distributions and run
update statistics for all tables and recompile all stored procedures afteran upgrade - that's the most common cause of performance degradation after
an upgrade because the internal data formats of the stored distributions
changes from version to version sometimes.
Art
Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
Blog: http://informix-myview.blogspot.com/
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.
On Wed, Sep 14, 2011 at 2:36 PM, Alberto Romeu Pessonio Filho <
arfilho@orizonbrasil.com.br> wrote:
> OK Art!!!
>
> But, in previous minor versions of IDS 11.50, when we apply new commercial
> versions (i.e: from 11.5 FC5 to 11.5 FC6) I noticed we had Page migrations.
> If we have page migrations, something is gonna change on a header structure
> of
> the Pages. In this case, we might cause decrease performance IF the
> maintenance hasn't a good analysis and plan. Although this change is easy
> to
> do, I need to have some precaution because it's a 24X7 core system business
> with a little time response to the application (billing authorization
> system).
> If something goes wrong, my head is gonna be hunted at my company.
>
> Anyway, I thank your suggestion, but it will be done in couple weeks.
>
> Thank you very much,
> Alberto.
>
> -----Mensagem original-----
> De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> Kagel
> Enviada em: quarta-feira, 14 de setembro de 2011 15:18
> Para: ids@iiug.org
> Assunto: Re: tempdbs used by a ghost [24925]
>
> Best practice:
>
> - Get 11.50.FC9 installed in a new directory during the day and copy in
>
> your ONCONFIG and sqlhosts files ready to go, and link any subdirectories
>
> that should be under the new INFORMIXDIR after the restart into the .FC9
>
> installation.
>
> - Then when you have to bounce your downtime will just be a minute or
>
> three to go down, relink INFORMIXDIR to point to the FC9 installation, and
>
> restart 11.50.FC9. Also if there is some problem starting .FC9, FC8 is
>
> still ready to go back online.
>
> - Go to sleep happy and comfortable.
>
> Art
>
> Art S. Kagel
> Advanced DataTools (www.advancedatatools.com)
> Blog: http://informix-myview.blogspot.com/
>
> Disclaimer: Please keep in mind that my own opinions are my own opinions
> and
> do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
> organization with which I am associated either explicitly, implicitly, or
> by
> inference. Neither do those opinions reflect those of other individuals
> affiliated with any entity with which I am affiliated nor those of the
> entities themselves.
>
> On Wed, Sep 14, 2011 at 1:58 PM, Alberto Romeu Pessonio Filho <
> arfilho@orizonbrasil.com.br> wrote:
>
> > Hi Art...
> >
> > I tried restart the PC of rrosini, but it not solved.
> > Unfortunately, i'll restart the engine late at night and apply 11.50 FC9
> > release, as Dave suggested. The situation is identical as the bug
> reports.
> >
> > Thank you guys!!!
> >
> > Best regards,
> > Alberto.
> >
> > -----Mensagem original-----
> > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> > Kagel
> > Enviada em: quarta-feira, 14 de setembro de 2011 13:06
> > Para: ids@iiug.org
> > Assunto: Re: tempdbs used by a ghost [24914]
> >
> > One thing to try first: If rrosini ran the query from a PC based tool,
> try
> > rebooting her PC first. That may release the network connection. PC apps
> > are terrible about properly shutting down their network connections and
> > sometimes that leaves Informix blocked waiting for a response from an app
> > that is long gone.
> >
> > Art
> >
> > Art S. Kagel
> > Advanced DataTools (www.advancedatatools.com)
> > Blog: http://informix-myview.blogspot.com/
> >
> > Disclaimer: Please keep in mind that my own opinions are my own opinions
> > and
> > do not reflect on my employer, Advanced DataTools, the IIUG, nor any
> other
> > organization with which I am associated either explicitly, implicitly, or
> > by
> > inference. Neither do those opinions reflect those of other individuals
> > affiliated with any entity with which I am affiliated nor those of the
> > entities themselves.
> >
> > On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho <
> > arfilho@orizonbrasil.com.br> wrote:
> >
> > > Hi Art!
> > >
> > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she
> > got
> > > the result yesterday. I think the query fetched data and finished.
> Also,
> > I
> > > saw
> > > the query, and this kind of query when is triggered on the instance
> > usually
> > > brings rows in 3 minutes, at max.
> > >
> > > I think that I really need to bounce the engine.
> > >
> > > Thanks in advance,
> > > Alberto.
> > >
> > > -----Mensagem original-----
> > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> > > Kagel
> > > Enviada em: quarta-feira, 14 de setembro de 2011 11:14
> > > Para: ids@iiug.org
> > > Assunto: Re: tempdbs used by a ghost [24908]
> > >
> > > If the rrosini session is already gone then there isn't anything you
> can
> > do
> > > to get rid of the orphaned temp space short of bouncing the instance.
> > > Usually it is still lying around because the sort is still going on
> > (could
> > > be a pathological query like a cartesian product sorting trillions of
> > rows)
> > > and the space will free once the sort completes and the engine tries to
> > > tell
> > > the session that data is finally available to fetch. Either that or it
> > will
> > > run out of temp dbspace and that will cause the query to rollback.
> Either
> > > way at that point it will realize that the session is gone and release
> > the
> > > resources the session is holding. However, if the query that is sorting
> > it
> > > VERY large that may not happen for quite some time.
> > >
> > > Art
> > >
> > > Art S. Kagel
> > > Advanced DataTools (www.advancedatatools.com)
> > > Blog: http://informix-myview.blogspot.com/
> > >
> > > Disclaimer: Please keep in mind that my own opinions are my own
> opinions
> > > and
> > > do not reflect on my employer, Advanced DataTools, the IIUG, nor any
> > other
> > > organization with which I am associated either explicitly, implicitly,
> or
> > > by
> > > inference. Neither do those opinions reflect those of other i
Art:
I appreciate your comment and your compliment.
Thank you all guys who helped me.
Regards,
Alberto.
-----Mensagem original-----
De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art Kagel
Enviada em: quarta-feira, 14 de setembro de 2011 15:54
Para: ids@iiug.org
Assunto: Re: tempdbs used by a ghost [24928]
Understandable and understood. The new features and structural changes in
11.50 stopped with 11.50.xC6 or 7 when 11.70 took over, so you should be
OK. That said, good DBAs are paranoid DBAs, so caution is the watchword and
planning too. It's always a good idea to run drop all distributions and run
update statistics for all tables and recompile all stored procedures afteran upgrade - that's the most common cause of performance degradation after
an upgrade because the internal data formats of the stored distributions
changes from version to version sometimes.
Art
Art S. Kagel
Advanced DataTools (www.advancedatatools.com)
Blog: http://informix-myview.blogspot.com/
Disclaimer: Please keep in mind that my own opinions are my own opinions and
do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
organization with which I am associated either explicitly, implicitly, or by
inference. Neither do those opinions reflect those of other individuals
affiliated with any entity with which I am affiliated nor those of the
entities themselves.
On Wed, Sep 14, 2011 at 2:36 PM, Alberto Romeu Pessonio Filho <
arfilho@orizonbrasil.com.br> wrote:
> OK Art!!!
>
> But, in previous minor versions of IDS 11.50, when we apply new commercial
> versions (i.e: from 11.5 FC5 to 11.5 FC6) I noticed we had Page migrations.
> If we have page migrations, something is gonna change on a header structure
> of
> the Pages. In this case, we might cause decrease performance IF the
> maintenance hasn't a good analysis and plan. Although this change is easy
> to
> do, I need to have some precaution because it's a 24X7 core system business
> with a little time response to the application (billing authorization
> system).
> If something goes wrong, my head is gonna be hunted at my company.
>
> Anyway, I thank your suggestion, but it will be done in couple weeks.
>
> Thank you very much,
> Alberto.
>
> -----Mensagem original-----
> De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> Kagel
> Enviada em: quarta-feira, 14 de setembro de 2011 15:18
> Para: ids@iiug.org
> Assunto: Re: tempdbs used by a ghost [24925]
>
> Best practice:
>
> - Get 11.50.FC9 installed in a new directory during the day and copy in
>
> your ONCONFIG and sqlhosts files ready to go, and link any subdirectories
>
> that should be under the new INFORMIXDIR after the restart into the .FC9
>
> installation.
>
> - Then when you have to bounce your downtime will just be a minute or
>
> three to go down, relink INFORMIXDIR to point to the FC9 installation, and
>
> restart 11.50.FC9. Also if there is some problem starting .FC9, FC8 is
>
> still ready to go back online.
>
> - Go to sleep happy and comfortable.
>
> Art
>
> Art S. Kagel
> Advanced DataTools (www.advancedatatools.com)
> Blog: http://informix-myview.blogspot.com/
>
> Disclaimer: Please keep in mind that my own opinions are my own opinions
> and
> do not reflect on my employer, Advanced DataTools, the IIUG, nor any other
> organization with which I am associated either explicitly, implicitly, or
> by
> inference. Neither do those opinions reflect those of other individuals
> affiliated with any entity with which I am affiliated nor those of the
> entities themselves.
>
> On Wed, Sep 14, 2011 at 1:58 PM, Alberto Romeu Pessonio Filho <
> arfilho@orizonbrasil.com.br> wrote:
>
> > Hi Art...
> >
> > I tried restart the PC of rrosini, but it not solved.
> > Unfortunately, i'll restart the engine late at night and apply 11.50 FC9
> > release, as Dave suggested. The situation is identical as the bug
> reports.
> >
> > Thank you guys!!!
> >
> > Best regards,
> > Alberto.
> >
> > -----Mensagem original-----
> > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> > Kagel
> > Enviada em: quarta-feira, 14 de setembro de 2011 13:06
> > Para: ids@iiug.org
> > Assunto: Re: tempdbs used by a ghost [24914]
> >
> > One thing to try first: If rrosini ran the query from a PC based tool,
> try
> > rebooting her PC first. That may release the network connection. PC apps
> > are terrible about properly shutting down their network connections and
> > sometimes that leaves Informix blocked waiting for a response from an app
> > that is long gone.
> >
> > Art
> >
> > Art S. Kagel
> > Advanced DataTools (www.advancedatatools.com)
> > Blog: http://informix-myview.blogspot.com/
> >
> > Disclaimer: Please keep in mind that my own opinions are my own opinions
> > and
> > do not reflect on my employer, Advanced DataTools, the IIUG, nor any
> other
> > organization with which I am associated either explicitly, implicitly, or
> > by
> > inference. Neither do those opinions reflect those of other individuals
> > affiliated with any entity with which I am affiliated nor those of the
> > entities themselves.
> >
> > On Wed, Sep 14, 2011 at 10:21 AM, Alberto Romeu Pessonio Filho <
> > arfilho@orizonbrasil.com.br> wrote:
> >
> > > Hi Art!
> > >
> > > I've talked to rrosini and she (Yes, a "dumb" girl...) told me that she
> > got
> > > the result yesterday. I think the query fetched data and finished.
> Also,
> > I
> > > saw
> > > the query, and this kind of query when is triggered on the instance
> > usually
> > > brings rows in 3 minutes, at max.
> > >
> > > I think that I really need to bounce the engine.
> > >
> > > Thanks in advance,
> > > Alberto.
> > >
> > > -----Mensagem original-----
> > > De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] Em nome de Art
> > > Kagel
> > > Enviada em: quarta-feira, 14 de setembro de 2011 11:14
> > > Para: ids@iiug.org
> > > Assunto: Re: tempdbs used by a ghost [24908]
> > >
> > > If the rrosini session is already gone then there isn't anything you
> can
> > do
> > > to get rid of the orphaned temp space short of bouncing the instance.
> > > Usually it is still lying around because the sort is still going on
> > (could
> > > be a pathological query like a cartesian product sorting trillions of
> > rows)
> > > and the space will free once the sort completes and the engine tries to
> > > tell
> > > the session that data is finally available to fetch. Either that or it
> > will
> > > run out of temp dbspace and that will cause the query to rollback.
> Either
> > > way at that point it will realize that the session is gone and release
> > the
> > > resources the session is holding. However, if the query that is sorting
> > it
> > > VERY large that may not happen for quite some time.
> > >
> > > Art
> > >
> > > Art S. Kagel
> > > Advanced DataTools (www.advancedatatools.com)
> > > Blog: http://informi