UNIX for Dummies Questions & Answers

Hi,



Recently my Fedora running under level 3 begun freezing. It was running for about four months without any failures, but now it freezes almost every day, even twice a day. Sometimes it comes back itself, sometimes hard reboot is needed.

This is a remote server where I connect through ssh. When freezed, can't connect to ssh, can't open http and asterisk which is running on this system is not accepting calls.

I couldn't find anything interesting in log files, only this when this system is loading:

Code:

Dec 25 21:01:18 srv1 kernel:  [<c042b0cf>] local_bh_enable+0x45/0x92
Dec 25 21:01:18 srv1 kernel:  [<c06002bd>] cond_resched_softirq+0x2c/0x42
Dec 25 21:01:18 srv1 kernel:  [<c05d2f54>] established_get_next+0x52/0xd1
Dec 25 21:01:18 srv1 kernel:  [<c05d4001>] tcp_seq_next+0x7c/0x88
Dec 25 21:01:18 srv1 kernel:  [<c048b3c2>] seq_read+0x17a/0x274
Dec 25 21:01:18 srv1 kernel:  [<c048b248>] seq_read+0x0/0x274
Dec 25 21:01:18 srv1 kernel:  [<c0475e85>] vfs_read+0xa6/0x152
Dec 25 21:01:18 srv1 kernel:  [<c04762db>] sys_read+0x41/0x67
Dec 25 21:01:18 srv1 kernel:  [<c0404f70>] syscall_call+0x7/0xb
Dec 25 21:01:18 srv1 kernel:  =======================

I am running Fedora 7 with kernel 2.6.21-1.3194.fc7
HDD: 250 GB (more than 90% free)
RAM: 1GB
CPU: Core 2 Duo 1.8GHz

I am just little familiar with unix system, so I need someone to help me to find out what is causing the system to freeze

Freezing almost every day doesn't sound like a software issue, but rather as hardware one. How many people have root access to the server, any changes been done recently, updates, patches ? Nothing suspicious in the dmesg ?

I had Asterisk running on this system with high priority. Recently changed priority to normal and no freeze so far.

I am only user with root access, no changes or updates were done recently before freezes. The only suspicios message is the message I've posted above.

I am also thinking of RAM or HDD problem. This is a remote system where I don't have physical access, besides it's a production system with more than 10 calls at any time during 24/7, so tests will be quite uncomfortable, but I am planning to do so after next failure.