Hey all,
I've been running into a few strange problems lately with a file server. I have around 15-20 other servers that are mapped to this system either via NFS or Samba running various scripts. The python scripts for these jobs are on the server, so each system is constantly reading the server. Some of these systems have a good 40-50 jobs running on them, and this setup has been working great for months.
Now, over the past week (since the last server reboot), random jobs have been freezing on the systems when trying to read or write to the file server. However, it's only 1-2 jobs that do this, and the other jobs on the system will continue working happily, with no problems reading from the server at all. Trying to browse to the mount point for the file server doesn't work after this happens either. Unmounting and remounting via command line also doesn't work, as things just hang. The only way I've been able to clear a system up when this happens is to reboot, and everything remounts just fine when booting up.
Most of these issues have happened on other Linux systems mounting with NFS, but there are a few Windows 2000 systems that have also had a random job freeze up lately and have needed to reboot to clear things up.
It seems to me that I'm hitting some kind of limit on the file server. Unfortunately, when this has happened, there are not any logs on either system that help in tracking down this problem.
I've done some searching, and have looking into problems caused by file descriptors and/or hitting the open files limit. I can't find anything with these symptoms though, so I'm hesitant to start modifying things like that without being sure.
The file server itself is fairly hefty. It's a Pentium D (dual core) system, 2GB of memory, SATAII RAID-5 array, etc. Fedora Core 5 is the operating system. Most other systems connecting to it are also on Fedora 5, with 1 or 2 still on Fedora 4. As mentioned above, we do have a few Windows 2000 Server systems connecting via Samba as well.
Any advice on this will really be appreciated.
Thanks!
I've been running into a few strange problems lately with a file server. I have around 15-20 other servers that are mapped to this system either via NFS or Samba running various scripts. The python scripts for these jobs are on the server, so each system is constantly reading the server. Some of these systems have a good 40-50 jobs running on them, and this setup has been working great for months.
Now, over the past week (since the last server reboot), random jobs have been freezing on the systems when trying to read or write to the file server. However, it's only 1-2 jobs that do this, and the other jobs on the system will continue working happily, with no problems reading from the server at all. Trying to browse to the mount point for the file server doesn't work after this happens either. Unmounting and remounting via command line also doesn't work, as things just hang. The only way I've been able to clear a system up when this happens is to reboot, and everything remounts just fine when booting up.
Most of these issues have happened on other Linux systems mounting with NFS, but there are a few Windows 2000 systems that have also had a random job freeze up lately and have needed to reboot to clear things up.
It seems to me that I'm hitting some kind of limit on the file server. Unfortunately, when this has happened, there are not any logs on either system that help in tracking down this problem.
I've done some searching, and have looking into problems caused by file descriptors and/or hitting the open files limit. I can't find anything with these symptoms though, so I'm hesitant to start modifying things like that without being sure.
The file server itself is fairly hefty. It's a Pentium D (dual core) system, 2GB of memory, SATAII RAID-5 array, etc. Fedora Core 5 is the operating system. Most other systems connecting to it are also on Fedora 5, with 1 or 2 still on Fedora 4. As mentioned above, we do have a few Windows 2000 Server systems connecting via Samba as well.
Any advice on this will really be appreciated.
Thanks!