> -----Original Message----- > From: redhat-list-bounces@xxxxxxxxxx [mailto:redhat-list- > bounces@xxxxxxxxxx] On Behalf Of Ben Russo > Sent: Tuesday, April 12, 2005 11:47 AM > To: General Red Hat Linux discussion list > Subject: Re: ps shows wierd clock time values ? > > Tobias Speckbacher wrote: > >>-----Original Message----- > >>From: redhat-list-bounces@xxxxxxxxxx [mailto:redhat-list- > >>bounces@xxxxxxxxxx] On Behalf Of Ben Russo > >>Sent: Tuesday, April 12, 2005 8:32 AM > >>To: redhat-list@xxxxxxxxxx > >>Subject: ps shows wierd clock time values ? > >> > >>Today I looked at a system that has weird timestamp/clock issues? > >>Anybody seen this before, know what it is? > >>I have some excerpts of commands below that demonstrate the issue. > >>Please note the 2006, and April 15 dates on some of the processes and > >>file timestamps as well as process times last year, when the "uptime" > >>says only 17 days (March 25th).... But I could swear that the box has > >>been up for a LONG time (like maybe 2+ years?) > >>Doesn't make sense. > > > > > > What you are seeing are symptoms of the "jiffies overflow" bug. > > The problem is that uptime and process related time values are stored as > > a 32 bit values which contain the number of jiffies elapsed (jiffie = > > 1/100th second). Basically the counter overflows after 497 days of > > uptime and the symptoms your system exhibits appear. > > > > Upgrading the kernel to a more recent version should fix the issue. > > > > > > > > Thanks, I just found that out myself... > > http://www.ussg.iu.edu/hypermail/linux/kernel/0202.2/0337.html > > But I am running a RedHat Enterprise Linux Kernel 2.4.9-e.27 > which was built on Tue 05 Aug 2003 by RedHat. > Isn't that kernel already patched for this OLD bug? > > I have tried searching RedHat bugzilla for the Jiffies bug to see > If I could find a specific kernel version that listed it as > "fixed" but I haven't seen that, just a mention that in the old > RedHat 7.1 it was "closed" > > If I upgrade the kernel to 2.4.9-e.59 (the latest RHEL-AS-2.1 kernel) > will that have the fix? I don't think this is fixed, I just checked include/linux/sched.h in the .59 source and it shows start_time typed as unsigned long which would mean it is still using 32 bit storage. Whereas the patches I found type this value to u64. Worth confirming with RH though if your support contract is still valid. > > -Ben. > > -- > redhat-list mailing list > unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe > https://www.redhat.com/mailman/listinfo/redhat-list -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
