Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a problem with my HP Proliant D320 G6 server, OS Debian Lenny.

From time to time, there is a kernel panic (which is not the issue here, I'm trying to find the anwer to it elsewhere, because I don't know the source) and the system hangs, obviously.

As I understand the ASR recovery - it should after defined period of time, when the system is unresponsive, restart the machine (first ask the system to restart, after no response perform the hard reset).

The problem is simple - after tonight kernel panic the server hanged for about four hours and no restart has been performed. I checked triple times that ASR is ENABLED and is set to 10 minutes check. I would like to be absolutely sure, that if this happens again, it will restart the machine. I cannot afford the production server to hang and wait for the telehouse to call me, that there is something wrong.

If someone has any advice, please share it. Thanks in advance.

share|improve this question
    
I can't comment on the ASR issue but I would recommend installing some monitoring system like Nagios on another machine. This allows you to get alerts nearly immediately if the server is down and also to check other parameters like hard disk usage, system load and the status of services. –  Sven Jul 10 '11 at 10:42
    
That's definitely weird, is the firmware up to date? I had a bizarre issue once, totally opposite of what you have - it kept ASR'ing a perfectly functioning machine! Updated the firmware in preparation for HP telling me to do that... didn't need to phone them after I updated the firmware though, as that seemed to fix it. –  Ben Pilbrow Jul 10 '11 at 10:50
    
@SvenW Monitoring software on another machine starts to look as a good solution. In present time I have no other server I could use. –  koressak Jul 13 '11 at 9:12
    
@Ben Pilbrow: I did almost every update I could (and was available), only the one instruction I could not do - turn it off and unplug the power source for 30 sec. But apart from that, the hp management software and firmware versions are up to date. –  koressak Jul 13 '11 at 9:14

2 Answers 2

I typically set the ASR timeout to 5 minutes... However, this is really intended to catch total lockups. In the case of a storage subsystem lockup, where I/O is blocked, the system load rises to 100+, BUT the server remains pingable, an ASR would not necessarily kick in. Also, you should be running the appropriate ProLiant Support Pack or management agents for your OS/hardware combination.

share|improve this answer
    
I'm running the HP support pack agents and as mentioned in my other comment, I did full firmware upgrade several months ago (the last update 3 weeks ago). And still the problem persists. –  koressak Jul 13 '11 at 9:17
    
You should not rely on ASR to maintain uptime or reduce downtime. You obviously have another issue with your system. Since you have agents installed, what does the HP IML output show? hplog -v Are there any errors? How about your ILO? Is it configured? What is on the screen when the kernel panics? –  ewwhite Jul 13 '11 at 13:49
    
Unfortunately HP log is clear of any event occuring on the system. Yes, the kernel panic is something that should be taken care of first. At the moment I haven't got a possibility to run ILO on it's own ip, because i don't have one spare. Kernel panic displays error during Machine check, Channel 5, Bank 4, TSF 0000. I'm bewildered by the zero error address. I found many kernel panic problems througout internet but nothing alike. –  koressak Jul 13 '11 at 18:29
    
That means that the RAM or motherboard has a problem. Memory channels and banks. The DL320 G6 has 3 channels and 9 memory slots. Please try reseating the RAM or running the memory diagnostic CD that came with the server. –  ewwhite Jul 13 '11 at 18:50
    
Ok. Will try that as soon as possible. –  koressak Jul 14 '11 at 12:41

ASR will only work if the service that runs on the server sets the timer at least once. If ASR isn't working when the computer completely locks up, and the watchdog service cannot reset the timer, then it must not have been set in the first place (the service must not be running).

share|improve this answer
    
Are you referring to the ASR daemon included with the agents? –  ewwhite Jul 13 '11 at 15:03
    
Yeah; I'm sure my terminology is a bit off in that answer... –  Chris S Jul 13 '11 at 16:54
    
I don't completely understand. I'm quite sure that every hp agent is working properly (if the system works fine), but after the kernel panic it stops responding to anything. –  koressak Jul 13 '11 at 18:31

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.