Re: reuse raw device

Hi, thanks :) I got that. When IDS instance is inited, it writes some information into the disk where ROOTPATH located.these information usually has something to do with OS, just like the result of "xxd /dev/mydiskdevice" below in Linux and windows: ... 00007f0: 1401 1c00 0000 0000 1800 fc00 6bfc 0000 ............k... 0000800: 0100 0000 0100 5200 5300 0018 a506 0b00 ......R.S....... 0000810: 0000 0000 0000 0000 524f 4f54 4e41 4d45 ........ROOTNAME 0000820: 2072 6f6f 7464 6273 0052 4f4f 5450 4154 rootdbs.ROOTPAT 0000830: 4820 2f68 6f6d 652f 696e 666f 726d 6978 H /home/informix 0000840: 2f64 6273 7061 6365 732f 726f 6f74 6462 /dbspaces/rootdb // in Linux 0000850: 7300 524f 4f54 4f46 4653 4554 2030 0052 s.ROOTOFFSET 0.R 0000860: 4f4f 5453 495a 4520 3230 3030 3030 004d OOTSIZE 200000.M 0000870: 4952 524f 5220 3000 4d49 5252 4f52 5041 IRROR 0.MIRRORPA 0000880: 5448 202f 6f70 742f 696e 666f 726d 6978 TH /opt/informix 0000890: 2f74 6d70 2f64 656d 6f5f 6f6e 2e72 6f6f /tmp/demo_on.roo ... in MSwindows : .... 0001000: 0100 0000 0100 4618 8000 0018 d00a 2c03 ......F.......,. 0001010: 0000 0000 0000 0000 524f 4f54 4e41 4d45 ........ROOTNAME 0001020: 2072 6f6f 7464 6273 0052 4f4f 5450 4154 rootdbs.ROOTPAT 0001030: 4820 5c5c 2e5c 443a 5c72 6f6f 7464 6273 H \\\\\\\\.\\\\D:\\\\rootdbs 0001040: 0052 4f4f 544f 4646 5345 5420 3000 524f .ROOTOFFSET 0.RO 0001050: 4f54 5349 5a45 2032 3030 3030 3000 4d49 OTSIZE 200000.MI 0001060: 5252 4f52 2030 004d 4952 524f 5250 4154 RROR 0.MIRRORPAT 0001070: 4820 004d 4952 524f 524f 4646 5345 5420 H .MIRROROFFSET // in Windows 0001080: 3000 4442 5f4c 4942 5241 5259 5f50 4154 0.DB_LIBRARY_PAT 0001090: 4820 004a 5650 484f 4d45 2043 3a5c 5052 H .JVPHOME C:\\\\PR 00010a0: 4f47 5241 7e31 5c49 424d 5c49 424d 494e OGRA~1\\\\IBM\\\\IBMIN 00010b0: 467e 315c 3131 2e35 305c 6578 7465 6e64 F~1\\\\11.50\\\\extend .... Most of these seems come from ONCONFIG file. Why they are here....? Just to compare with the local ONCONFIG file? if anything is different, error comes... Another question, what does this msg below indicate? 19:14:28 Assert Warning: chunk failed sanity check 19:14:28 IBM Informix Dynamic Server Version 11.50.FC4DE 19:14:28 Who: Session(1, informix@systemp5, 0, 70000001036f028) Thread(7, main_loop(), 70000001032d028, 1) File: rspartn.c Line: 9350 19:14:28 Results: Chunk 1 is being taken OFFLINE. 19:14:28 Action: For all spaces other than temporary dbspaces, restore the space containing the chunk from the archive. If this chunk belongs to a temporary dbspace and if it is not the only chunk in that dbspace, then drop and re-create the chunk. Otherwise drop and re-create the temporary dbspace. 19:14:28 stack trace for pid 639034 written to /home/db/tmp/af.3ef6904 19:14:28 See Also: /home/db/tmp/af.3ef6904 19:14:28 chunk failed sanity check 19:14:29 I/O error, Primary Chunk '/home/informix/dbspaces/shared.dsik1' -- Offline (sanity) 19:14:29 oninit: Fatal error in shared memory initialization jc 2009/9/15 tilleul17@web.de <tilleul17@web.de> > wrote: > > Hi, > > > > the endian problem is clear, but the OS problem is not. > > > > When we use raw device, there is nothing to do with filesystem. Our > > read/write operation on the raw disk is done by IDS itself. > > So in IDS 11.5, the page size of new dbspace is configurable when it is > > created. But how to configure the page size of the dbspace in ROOTPATH? > > > > if the page size of dbspaces are of the same, including dbspace in > ROOTPATH, > > Can that work? > > > > Is there any other difference of IDS on different OS? > > > > thanks :) > > > > jc. > > > > 2009/9/10 Art S. Kagel <art.kagel@gmail.com> > > > >> No. That will not work except accidentally. Certainly it is officially > un= > >> supported. > >> > >> Art=20 > > I reiterate Art - its not supported. > > As said it works under certain conditions and - to my knowledge - has > been done successfully before ( RELIANT to SOLARIS if I recall it > correctly ) > > You would have to take care of locales - if its a legacy system and > there are databases that got created with OS locales you might run into > all kinds of troubles - if the locales on one platform are somewhat > different from the source platform. > This is one caveat to be aware of. Others might be there. > > As for ROOTPATH - you don't configure any critical dbspace for larger > pagesize, that#s fixed to 2 k (4K on AIX and Windows) > > But again for the reason that it is not supported - don't do it. > > HTH > Tilman > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --0016e64cc54c720a1204739cc0ea

The zeroth page in the first rootdbs chunk identifies the OS and IDS version and wordsize as well as the root chunk pagesize and other information that the starting engine needs to know if the chunk is the correct chunk, if there are some version compatibility problems, and if it needs to perform a version upgrade. The following pages contain the lists of dbspaces and chunks and their mirrors, records of archives and the most recent checkpoints, and a copy of the ONCONFIG file that was current the last time the engine was started. This is so that the engine can note in the online log what parameters changed between the previous startup/shutdown cycle and the current startup. Also, in earlier versions, some disk modifications were driven by these configuration changes (such as changing PHYSDBS or PHYSFILE which you can no longer do that way). Art Art S. Kagel Oninit (www.oninit.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Oninit, the IIUG, nor any other organization with which I am associated either explicitly or implicitly. 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 Tue, Sep 15, 2009 at 8:16 AM, <jencce2002@gmail.com> wrote: > Hi, > > thanks :) > > I got that. > When IDS instance is inited, it writes some information into the disk where > ROOTPATH located.> these information usually has something to do with OS, just like the result > of "xxd /dev/mydiskdevice" below in Linux and windows: > .... > 00007f0: 1401 1c00 0000 0000 1800 fc00 6bfc 0000 ............k... > 0000800: 0100 0000 0100 5200 5300 0018 a506 0b00 ......R.S....... > 0000810: 0000 0000 0000 0000 524f 4f54 4e41 4d45 ........ROOTNAME > 0000820: 2072 6f6f 7464 6273 0052 4f4f 5450 4154 rootdbs.ROOTPAT > 0000830: 4820 2f68 6f6d 652f 696e 666f 726d 6978 H /home/informix > 0000840: 2f64 6273 7061 6365 732f 726f 6f74 6462 > /dbspaces/rootdb // in Linux > 0000850: 7300 524f 4f54 4f46 4653 4554 2030 0052 s.ROOTOFFSET 0.R > 0000860: 4f4f 5453 495a 4520 3230 3030 3030 004d OOTSIZE 200000.M > 0000870: 4952 524f 5220 3000 4d49 5252 4f52 5041 IRROR 0.MIRRORPA > 0000880: 5448 202f 6f70 742f 696e 666f 726d 6978 TH /opt/informix > 0000890: 2f74 6d70 2f64 656d 6f5f 6f6e 2e72 6f6f /tmp/demo_on.roo > .... > in MSwindows : > ..... > 0001000: 0100 0000 0100 4618 8000 0018 d00a 2c03 ......F.......,. > 0001010: 0000 0000 0000 0000 524f 4f54 4e41 4d45 ........ROOTNAME > 0001020: 2072 6f6f 7464 6273 0052 4f4f 5450 4154 rootdbs.ROOTPAT > 0001030: 4820 5c5c 2e5c 443a 5c72 6f6f 7464 6273 H \\\\\\\\.\\\\D:\\\\rootdbs > 0001040: 0052 4f4f 544f 4646 5345 5420 3000 524f .ROOTOFFSET 0.RO > 0001050: 4f54 5349 5a45 2032 3030 3030 3000 4d49 OTSIZE 200000.MI > 0001060: 5252 4f52 2030 004d 4952 524f 5250 4154 RROR 0.MIRRORPAT > 0001070: 4820 004d 4952 524f 524f 4646 5345 5420 H > ..MIRROROFFSET // in Windows > 0001080: 3000 4442 5f4c 4942 5241 5259 5f50 4154 0.DB_LIBRARY_PAT > 0001090: 4820 004a 5650 484f 4d45 2043 3a5c 5052 H .JVPHOME C:\\\\PR > 00010a0: 4f47 5241 7e31 5c49 424d 5c49 424d 494e OGRA~1\\\\IBM\\\\IBMIN > 00010b0: 467e 315c 3131 2e35 305c 6578 7465 6e64 F~1\\\\11.50\\\\extend > ..... > > Most of these seems come from ONCONFIG file. > > Why they are here....? Just to compare with the local ONCONFIG file? if > anything is different, error comes... > > Another question, what does this msg below indicate? > > 19:14:28 Assert Warning: chunk failed sanity check > 19:14:28 IBM Informix Dynamic Server Version 11.50.FC4DE > 19:14:28 Who: Session(1, informix@systemp5, 0, 70000001036f028) > > Thread(7, main_loop(), 70000001032d028, 1) > > File: rspartn.c Line: 9350 > 19:14:28 Results: Chunk 1 is being taken OFFLINE. > 19:14:28 Action: For all spaces other than temporary dbspaces, restore the > space > > containing the chunk from the archive. If this chunk belongs to a > > temporary dbspace and if it is not the only chunk in that > dbspace, > > then drop and re-create the chunk. Otherwise drop and re-create > > the temporary dbspace. > 19:14:28 stack trace for pid 639034 written to /home/db/tmp/af.3ef6904 > 19:14:28 See Also: /home/db/tmp/af.3ef6904 > 19:14:28 chunk failed sanity check > 19:14:29 I/O error, Primary Chunk '/home/informix/dbspaces/shared.dsik1' -- > Offline (sanity) > 19:14:29 oninit: Fatal error in shared memory initialization > > jc > > 2009/9/15 tilleul17@web.de <tilleul17@web.de> > > > wrote: > > > Hi, > > > > > > the endian problem is clear, but the OS problem is not. > > > > > > When we use raw device, there is nothing to do with filesystem. Our > > > read/write operation on the raw disk is done by IDS itself. > > > So in IDS 11.5, the page size of new dbspace is configurable when it is > > > created. But how to configure the page size of the dbspace in ROOTPATH? > > > > > > if the page size of dbspaces are of the same, including dbspace in > > ROOTPATH, > > > Can that work? > > > > > > Is there any other difference of IDS on different OS? > > > > > > thanks :) > > > > > > jc. > > > > > > 2009/9/10 Art S. Kagel <art.kagel@gmail.com> > > > > > >> No. That will not work except accidentally. Certainly it is officially > > un= > > >> supported. > > >> > > >> Art=20 > > > > I reiterate Art - its not supported. > > > > As said it works under certain conditions and - to my knowledge - has > > been done successfully before ( RELIANT to SOLARIS if I recall it > > correctly ) > > > > You would have to take care of locales - if its a legacy system and > > there are databases that got created with OS locales you might run into > > all kinds of troubles - if the locales on one platform are somewhat > > different from the source platform. > > This is one caveat to be aware of. Others might be there. > > > > As for ROOTPATH - you don't configure any critical dbspace for larger > > pagesize, that#s fixed to 2 k (4K on AIX and Windows) > > > > But again for the reason that it is not supported - don't do it. > > > > HTH > > Tilman > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > --0016e64cc54c720a1204739cc0ea > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --00151747ba9cffdb7f04739f5cd7

wrote: ... > When IDS instance is inited, it writes some information into the disk where > ROOTPATH located.> these information usually has something to do with OS, just like the result > of "xxd /dev/mydiskdevice" below in Linux and windows: > .... > 00007f0: 1401 1c00 0000 0000 1800 fc00 6bfc 0000 ............k... > 0000800: 0100 0000 0100 5200 5300 0018 a506 0b00 ......R.S....... .... > ..... > > Most of these seems come from ONCONFIG file. > > Why they are here....? Just to compare with the local ONCONFIG file? if > anything is different, error comes... It seems to me you are a hacker :-) These are the ROOT RESERVED PAGES, kind of boot sector for IDS. If you are really interested in these details - and I think any DBA should understand IDS to that level of detail - I suggest that you take the training IDS Server Administration Part II. > > Another question, what does this msg below indicate? > > 19:14:28 Assert Warning: chunk failed sanity check > 19:14:28 IBM Informix Dynamic Server Version 11.50.FC4DE > 19:14:28 Who: Session(1, informix@systemp5, 0, 70000001036f028) > > Thread(7, main_loop(), 70000001032d028, 1) > > File: rspartn.c Line: 9350 > 19:14:28 Results: Chunk 1 is being taken OFFLINE. It means that you ROOT Chunk is corrupt - i.e. IDS instance need to be restored from a valid backup. Look into the generated file /home/db/tmp/af.3ef6904 it might give you more info what the nature of the corruption is . Regards Tilman