Cooked/Raw deivce in the same machine??

Our informix is running on an unix machine having 2 hard disks. We use 1 hard disk as a raw device for informix. The problem is that there is no enough space in that raw device. I am thinking about creating a cooked device on the other hard disk and increasing database space. Is it possible to use raw device and cooked device concurrently in the same machine ? Do we have to execute 2 engines?(I don't want this.) If possible, is it stable? Maybe, we can change the raw device to cooked device But, how about the performance? If you had a similar experience, please help me! Do you need version information? Informix 7.14 on HP-UX 10.01. Sorry for poor English and poor DBMS knowledge. everybody... Be Happy!! :-)

In article <E3T_3.673$sI3.25399@news.hananet.net>, '迵'' <ninezero@chollian.net> wrote: >I am thinking >about creating a cooked device on the other hard disk >and increasing database space. >Is it possible to use raw device and cooked device >concurrently in the same machine ? >Do we have to execute 2 engines?(I don't want this.) >If possible, is it stable? >Maybe, we can change the raw device to cooked device >But, how about the performance? Yes, it's possible, it's stable, and you only need the one instance. Differences in performance depend on the OS; you'll likely see a small degradation on HP-UX. Cooked space is buffered in two places: the Informix buffer cache, and the Unix file system buffer cache. The bigger down-side to using cooked space is that your data is not committed to disk until the syncer flushes the Unix cache. This can impact data integrity. If you have tables that don't change often, put them in their own dbspace on the cooked file, and you should be fine. Cheers, Jim -- W. Jim Jordan, Nortel Networks, Stop 29CA3A08 | +1 613 763 1568 PO Box 3511 Stn C, Ottawa, ON K1Y 4H7 Canada | wjjordan@nortelnetworks.com Opinions expressed are not necessarily those of Nortel Networks. Inbound spam filtering is in place. Don't send what I won't see.

There should be no problems, It is just that you may find I/O slowness in cooked device and not to forget somebody might accidentely knock off cooked file. (BEWARE!!!) Rgds "±è¿µ±Ô" wrote: > Our informix is running on an unix machine > having 2 hard disks. > > We use 1 hard disk as a raw device for informix. > > The problem is that there is no enough space in that raw device. > > I am thinking > about creating a cooked device on the other hard disk > and increasing database space. > > Is it possible to use raw device and cooked device > concurrently in the same machine ? > Do we have to execute 2 engines?(I don't want this.) > If possible, is it stable? > Maybe, we can change the raw device to cooked device > But, how about the performance? > > If you had a similar experience, > please help me! > > Do you need version information? > Informix 7.14 on HP-UX 10.01. > > Sorry for poor English and poor DBMS knowledge. > > everybody... Be Happy!! > :-)

In article <E3T_3.673$sI3.25399@news.hananet.net>, "'迵''" <ninezero@chollian.net> wrote: > > Our informix is running on an unix machine > having 2 hard disks. > > We use 1 hard disk as a raw device for informix. > > The problem is that there is no enough space in that raw device. > > I am thinking > about creating a cooked device on the other hard disk > and increasing database space. No problem, just be sure to use a symbolic link for the chunk path so that if you purchase another disk in the future you will be able to dd the cooked file to the new RAW device, point the link to the new partition RAW devicename, and restart the engine to go all RAW again. > Is it possible to use raw device and cooked device > concurrently in the same machine ? Yes. > Do we have to execute 2 engines?(I don't want this.) No. > If possible, is it stable? Yes. > Maybe, we can change the raw device to cooked device IFF you used a link for the chunk path it is possible but not desireable. You would use the method I described above for replacing the COOKED file with a RAW device later. > But, how about the performance? Cooked is 15-35% slower than RAW. Art S. Kagel Sent via Deja.com http://www.deja.com/ Before you buy.

Hi, as I know when Informix used kernel IO anywhere it will try first a kernel IO anywhere and then if it not possible (cooked file) it will do it with an ansync. IO. So that the cooked files will be a little bit slower as an only cokked file database Michael Chetan Anand schrieb: > There should be no problems, It is just that you may find I/O slowness > in cooked device and not to forget somebody might accidentely knock off > cooked file. (BEWARE!!!) > > Rgds > > "±è¿µ±Ô" wrote: > > > Our informix is running on an unix machine > > having 2 hard disks. > > > > We use 1 hard disk as a raw device for informix. > > > > The problem is that there is no enough space in that raw device. > > > > I am thinking > > about creating a cooked device on the other hard disk > > and increasing database space. > > > > Is it possible to use raw device and cooked device > > concurrently in the same machine ? > > Do we have to execute 2 engines?(I don't want this.) > > If possible, is it stable? > > Maybe, we can change the raw device to cooked device > > But, how about the performance? > > > > If you had a similar experience, > > please help me! > > > > Do you need version information? > > Informix 7.14 on HP-UX 10.01. > > > > Sorry for poor English and poor DBMS knowledge. > > > > everybody... Be Happy!! > > :-)