Memory Block Header corruption

I'm having a weird problem. Is this a known bug, and if so, is there a fix? A user attempts to log in to the database with an incorrect password. (This may not be related, but it's happened every time the database has crashed, so I included it.) Between 4 and 15 minutes later, the following message shows up in the log: Assert Failed: Memory block header corruption detected in mt_shm_free 1Then the database dies. Informix 9.14.UC7X4 running on Sun 5.6 12GB Memory, 12 processors. The database is not configured well, but I haven't had time to do anything about it. Nothing dangerous, just a poor performer right now. -- # unrm / ksh: unrm: not found # man cpio Sent via Deja.com http://www.deja.com/

In article <95n0r8$dkg$1@nnrp1.deja.com>, mars1972@my-deja.com writes >I'm having a weird problem. Is this a known bug, and if so, is there a >fix? > >A user attempts to log in to the database with an incorrect password. >(This may not be related, but it's happened every time the database has >crashed, so I included it.) >Between 4 and 15 minutes later, the following message shows up in the >log: >Assert Failed: Memory block header corruption detected in mt_shm_free 1>Then the database dies. > What does the /tmp/af. file say? >Informix 9.14.UC7X4 running on Sun 5.6 >12GB Memory, 12 processors. The database is not configured well, but I >haven't had time to do anything about it. Nothing dangerous, just a >poor performer right now. > > >-- ># unrm / >ksh: unrm: not found ># man cpio > > >Sent via Deja.com >http://www.deja.com/ -- David Williams

In article <4v5v5FBki0f6Ews2@smooth1.demon.co.uk>, David Williams <djw@smooth1.demon.co.uk> wrote: > In article <95n0r8$dkg$1@nnrp1.deja.com>, mars1972@my-deja.com writes > >I'm having a weird problem. Is this a known bug, and if so, is there a > >fix? > > > >A user attempts to log in to the database with an incorrect password. > >(This may not be related, but it's happened every time the database has > >crashed, so I included it.) > >Between 4 and 15 minutes later, the following message shows up in the > >log: > >Assert Failed: Memory block header corruption detected in mt_shm_free 1> >Then the database dies. > > > What does the /tmp/af. file say? > Not much that I can find that's useful. I'd send it to you, but it's half a meg. Is there something specific you were looking for? Or do you have an email address that I can send it to you as an attachment? The problem with the af file is that the problem appears to be caused by a process other than the one that actually brings the system down. It's like the first process corrupts some memory, but the database doesn't use that memory again until another process connects, so the process that caused the problem isn't the one that's reported. > >Informix 9.14.UC7X4 running on Sun 5.6 > >12GB Memory, 12 processors. The database is not configured well, but I > >haven't had time to do anything about it. Nothing dangerous, just a > >poor performer right now. > > > > > >-- > ># unrm / > >ksh: unrm: not found > ># man cpio > > > > > >Sent via Deja.com > >http://www.deja.com/ > > -- > David Williams > -- # unrm / ksh: unrm: not found # man cpio Sent via Deja.com http://www.deja.com/

In article <95p83h$9oq$1@nnrp1.deja.com>, mars1972@my-deja.com writes >In article <4v5v5FBki0f6Ews2@smooth1.demon.co.uk>, > David Williams <djw@smooth1.demon.co.uk> wrote: >> In article <95n0r8$dkg$1@nnrp1.deja.com>, mars1972@my-deja.com writes >> >I'm having a weird problem. Is this a known bug, and if so, is there >a >> >fix? >> > >> >A user attempts to log in to the database with an incorrect password. >> >(This may not be related, but it's happened every time the database >has >> >crashed, so I included it.) >> >Between 4 and 15 minutes later, the following message shows up in the >> >log: >> >Assert Failed: Memory block header corruption detected in >mt_shm_free 1>> >Then the database dies. >> > >> What does the /tmp/af. file say? >> >Not much that I can find that's useful. I'd send it to you, but it's >half a meg. Is there something specific you were looking for? Or do >you have an email address that I can send it to you as an attachment? > The stack trace from the thread which crashed.. >The problem with the af file is that the problem appears to be caused by >a process other than the one that actually brings the system down. It's >like the first process corrupts some memory, but the database doesn't >use that memory again until another process connects, so the process >that caused the problem isn't the one that's reported. > >> >Informix 9.14.UC7X4 running on Sun 5.6 >> >12GB Memory, 12 processors. The database is not configured well, but >I >> >haven't had time to do anything about it. Nothing dangerous, just a >> >poor performer right now. >> > >> > >> >-- >> ># unrm / >> >ksh: unrm: not found >> ># man cpio >> > >> > >> >Sent via Deja.com >> >http://www.deja.com/ >> >> -- >> David Williams >> > >-- ># unrm / >ksh: unrm: not found ># man cpio > > >Sent via Deja.com >http://www.deja.com/ -- David Williams