Red Hat Bugzilla – Bug 177030
ia64 floating-point assist fault from ntpd
Last modified: 2007-11-30 17:07:22 EST
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.10) Gecko/20050719 Red Hat/1.0.6-1.4.1 Firefox/1.0.6 Description of problem: When running for more than a couple of days I start seeing the following messages on my ia64 systems. I am seeing this on both RHEL4 U2 and U3 so it is not a (recent) regression. ntpd(2355): floating-point assist fault at ip 2000000800047962, isr 0000040000000008 ntpd(2355): floating-point assist fault at ip 2000000800047971, isr 0000020000000008 ntpd(2355): floating-point assist fault at ip 20000008001eb701, isr 0000020000000008 these 3 messages all happen at the same time in approx 1 minute intervals (which fills up logs pretty quickly). The ntpd binary is stripped so I am not able to determine what part of the code is causing this. I will try to build a debug version and see if I can reproduce with that. Since it sounds like other people have not seen this it might be something due to my ntpd.conf file. I will attach that here. Version-Release number of selected component (if applicable): ntp-4.2.0.a.20040617-4 How reproducible: Always Steps to Reproduce: 1. Install my ntpd.conf on an ia64 system 2. start ntpd 3. wait a couple of days Additional info:
Created attachment 122831 [details] ntp.conf file (I believe this is unmodified from the rpm but just in case...) Attaching the ntp.conf file. Nothing special and I don't think I have modified it from what was supplied in the rpm.
Created attachment 122871 [details] ntp-4.2.0.a.20050816-2.src.rpm Thanks for your ntp.conf file. ntpd is running on our common x86_64 machine (using your ntp.conf) now. Can you try it too (in your scenario, using *.src.rpm above) ? It creates /tmp/ntpd.PID and /tmp/ntpd-last.PID files. Please, can you send me these files (after 'crash') ? /tmp/ntpd.PID (ca. first 10000 lines) /tmp/ntpd-last.PID (ca. last 1000 lines)
I am running using the .src.rpm now. I will let it run over the weekend since it seems to take a day or two before the problem shows up. FYI, I don't think you will see the behavior on x86_64 since I _think_ the "floating-point assist" message is an ia64 specific thing. Also, for clarification on your request. You say to send the files after "crash". Are you saying this version of ntp should crash ntp when I get the errors? (it was not crashing before, just kept spitting out the messages". Also, not sure what you mean by "ca. first 10000 lines".
*** This bug has been marked as a duplicate of 147980 ***