Re: informix on Linux 2.4.1

Yes , I have succeeded Using raw device on Redhat 7.0 . because many engineers says Redhat 7.0 is not stable , they recommend Using Redhat 6.2 . thus, I upgrade kernel 2.4 from kernel2.2.16(Redhat6.2) . of course , i read shared memory bug in kernel archive. at kernel 2.2.16 and kernel 2.4 many many trials and errors , in shared memory allocation to informix , now I allocate about 2Gbytes shared memory to INFORMIX . till now i have no problem on this setting. but need to do more and more tests about this setting's satblity . "Sergey E. Volkov" <sve@raiden.bancorp.ru> wrote in message news:9603lp$991$1@nnrp1.deja.com... > Hi, > > There are no errors in the shared memory subsystem on 2.4.1 ( only in > 2.4.0 ), so you can to use raw device with 2.4.1 without any trouble. > > Good luck. > > Sergey. > > In article <3A82F5B9.B6381555@pandora.be>, > David Stes <stes@pandora.be> wrote: > > > > Hi, > > > > I've upgraded my machine to kernel 2.4.1 and I wonder whether I could > > now have 'raw devices' on Linux. > > > > I believe that on Linux 2.2 you had to use a patch from SGI or RedHat > > for the raw device support. > > > > Is this in kernel 2.4.1 still an issue ? Is anyone running IDS 9.21 > with > > raw devices on Linux kernel 2.4.1 ? > > > > Thanks! > > david. > > > > > Sent via Deja.com > http://www.deja.com/

Well, But there are same errors in 2.4.1 code, so be careful ... I think it's better way to wait for 2.4.2 ( or to using 2.4.1-pre2 ? ) to using on production environment. "But what's interesting here are actually three very small patches: - mm/memory.c (PageReserved() check): this could corrupt the shared zero page if you used direct IO, and thus make your system unusable - elevator fix (another missed re-initialization in __make_request(), and this time we made sure to check that everything else _is_ initialized) - IDE driver multi-mode write fix (ide_multwrite()). The first would only hit you if you used raw IO (and had some unlucky timing etc), and very few people do. The second can cause disk corruption with pretty much any disk (seen at least on SCSI under heavy load). Not necessarily easy to trigger, but still.. The third can cause disk corruption on IDE disks if you are using PIO writes with multi-mode and irq unmasking enabled. " Linus said. Goods luck. Sergey. In article <7lXxJAnkAHA.321@news1.sys.netsgo.com>, "jayhawks" <jayhawks@netsgo.com> wrote: > Yes , I have succeeded Using raw device on Redhat 7.0 . > because many engineers says Redhat 7.0 is not stable , > they recommend Using Redhat 6.2 . > > thus, I upgrade kernel 2.4 from kernel2.2.16(Redhat6.2) . > > of course , i read shared memory bug in kernel archive. > > at kernel 2.2.16 and kernel 2.4 many many trials and errors , > in shared memory allocation to informix , > now I allocate about 2Gbytes shared memory to INFORMIX . > > till now i have no problem on this setting. > but need to do more and more tests about this setting's satblity . > > "Sergey E. Volkov" <sve@raiden.bancorp.ru> wrote in message > news:9603lp$991$1@nnrp1.deja.com... > > Hi, > > > > There are no errors in the shared memory subsystem on 2.4.1 ( only in > > 2.4.0 ), so you can to use raw device with 2.4.1 without any trouble. > > > > Good luck. > > > > Sergey. > > > > In article <3A82F5B9.B6381555@pandora.be>, > > David Stes <stes@pandora.be> wrote: > > > > > > Hi, > > > > > > I've upgraded my machine to kernel 2.4.1 and I wonder whether I could > > > now have 'raw devices' on Linux. > > > > > > I believe that on Linux 2.2 you had to use a patch from SGI or RedHat > > > for the raw device support. > > > > > > Is this in kernel 2.4.1 still an issue ? Is anyone running IDS 9.21 > > with > > > raw devices on Linux kernel 2.4.1 ? > > > > > > Thanks! > > > david. > > > > > > > > > Sent via Deja.com > > http://www.deja.com/ > > Sent via Deja.com http://www.deja.com/

"Sergey E. Volkov" wrote: > > Well, > > But there are same errors in 2.4.1 code, so be careful ... > Yes you're right. I had a machine crash. But after recovering from that, now it seems to work. Here's what I did : * installed Slackware 7.1 * recompiled linux kernel 2.4.1 without doing anything special (the CONFIG_RAW is gone it seems, not needed any longer) * I downloaded 'raw' utility from ftp://ftp.uk.linux.org:/pub/linux/sct/fs/raw-io/ then I ran 'raw' to create bindings from /dev/raw1 and /dev/raw2 and /dev/raw3. then I configured informix to used a MIRROR'ed rootdbs on /dev/raw1 and /dev/raw2 (actually symbolic links to those) and I want to move my logical logs to /dev/raw3 dbs later. then I say a little prayer and type 'oninit -i' Wow ... seems to work. Until ... unfortunately my machine locks up. It was busy doing the sysmaster tables. I can't do anything any longer with my machine ! Help! So I power off and reboot. I start IDS with oninit (actually thinking that this would NOT work since I interrupted the oninit -i). Informix is going into fast recovery mode. It rebuilds indexes. The system is online. Strange. Well, it seems to work. I've been running the test database on this. I have now a mirrored root dbs on the raw device partitions, but I wonder why I had this crash.