System Dual P-PRO 200 Tyan
2 BusLogic BT-948's and one md0 device using raid0
consisting of 6 Seagate 4GB Hawks.
Application: multiple full news feeders and readers..
I've been testing new News software has lead to almost a crash
a day and fsck cleans up /dev/md0 perfectly with no
file corruption noticeable. Have been running this configuration
with INN for about a 6 mos and have yet to lose any data (knock on wood)
Regards
On Sun, 8 Mar 1998, Isaac Connor wrote:
>
> I have a question about the currnt stability of MD RAID0. The reason I
> ask is that since implementing it, I have experienced serious
> instability, and finaly complete utter an total filesystem corruption in
> the space of 48 hours uptime. (Fresh install to complete crap in 48
> hours). I am trying to determine whether the md code is at fault, or
> perhaps if using it has revealed a problem with my hardware. I am just
> curious if others use this and find it stable. Here is my system:
>
> Tyan Tomcat IIID
> 2 p200
> 48M EDO ram.
> Buslogic bt958 SCSI
> Glibc 206 based system, 2.1.8* kernels (I tried several).
>
> md setup
> I have an ide 2.0Gb IDE Western Digital Caviar series HD on hdb
> I have a Seagate 2.1Gb SCSI HD on sda
>
> Using e2fsprogs 1.10, and I tried the WIP 1.12 version.
>
> The problems I have are:
> random hard lock. This occurs more often under small load, and serious
> stress tests do not even touch the system, it grinds through the worst
> load I can give it (several make -j compile, X, Netscape, millions of
> xterms, playing mp3s, etc.
> Filesystem corruption. The seriousness and extent of corruption is
> proportional to the uptime. Even no load causes corruption. From a
> fresh slackware install, boot, build of raidtools, creation of md0,
> moving usr and home to the raid0 disk, reboot, and 5 minutes of uptime,
> upon the next reboot, there is serious corruption, and upon fscking
> manually, there are files in lost+found that are screwed. Undeelteable,
> modes of 12312312, sizes over terabytes and as small as 0bytes. Device
> files modes using almost all letters of the alphabet, you get the
> picture.
>
> So you get the idea, serious problems. This machine has been stable
> under various 2.1 kernels for about a year now, but in the last 2 weeks,
> it is nearly useless. Diagnostics on my hardware turn up nothing, but I
> am tempted to think that it is either hardware or MD. Please give me
> your thoughts.
>
> BTW, through all of this there is no debugging info in the logs, etc.
> The hard locks I mentioned leave no trace. They are running happily, and
> then the machine just stops, no ping, etc. This is all under X, XFree
> 3.3.1 compile for glibc.
>
> I can't think of anything else to give for info. Thank you for spending
> time reading this.
>
> Isaac Connor
> iconnor@uwaterloo.ca
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu