System hang during write of large files

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hello Janne,
nice to not be alone ...

We did some things during the last accident:
- Changed the alive time of buffer to 3 seconds from 30
- Started the wakeup of the kupdate daemon every 0,6 seconds
These both things are described in
http://www-106.ibm.com/developerworks/library/l-fs8
(search for the work kupdate)

We also changed from data=ordered to data=journal (seems to be better 
for database usage).

We did not tested since the changes where made because some persons are 
working on the machine (and after working a few weekends in a line I got 
a little bit tired).

Annother try worth in my eyes would be a bigger journal. Our current 
journal is 33 MB on a 40 and a 60 MB disk. If its a journal overflow may 
be 500 MB will solve the problem.

As I understand from your mail it is not a kernel problem of kernel 2.4.10.

Klaus


Janne Pikkarainen wrote:
>>    As we tried 3 of these extents in one run of the extent utility, 
>> the system hangs without any error message. The master console was 
>> also dead; we could enter root but the password prompt never comes up.
> 
> 
> Wow. Sounds just like the problem we currently have with one of our servers
> running Red Hat 7.2. It's hard to say if it's really an ext3-related 
> problem but anyway the server freezes whenever it wants to do it. The 
> same symptoms than above - login prompt allows me to enter root but the 
> password prompt never appears. If I'm already logged in, I may be able 
> to use commands which are already loaded in memory. Stuff which has to 
> be loaded from hd never loads.
> 
> Physically the server seems to be blinking it hard drive lights as 
> normally.
> 
> And the server...
> 
> - IBM xSeries 330
> - Red Hat 7.2
> - 1 GB RAM
> - 200 GB HD w/ ext3
> - IBM ServeRAID 4M
> - 2x Adaptec 7899P
> 
> The latest Red Hat cooked kernel I tried (2.4.9-21) gives something like 
> 1-6 hours of uptime. My own home-cooked and salted kernels are able to 
> run our server something like 10-40 days before crashing, no more. At 
> least not yet. :-) The latest kernel I have compiled myself is 
> 2.4.18-pre9 and it's having the same symptoms...
> 
> 
> 
> 
> _______________________________________________
> 
> Ext3-users@redhat.com
> https://listman.redhat.com/mailman/listinfo/ext3-users
> 







[Index of Archives]         [Linux RAID]     [Kernel Development]     [Red Hat Install]     [Video 4 Linux]     [Postgresql]     [Fedora]     [Gimp]     [Yosemite News]

  Powered by Linux