RE: ataraid + geom_stripe problems

From: Daniel Eriksson (daniel_k_eriksson_at_telia.com)
Date: 08/30/04

  • Next message: Maxim Sobolev: "burncd(8) usability: why `-s max' isn't default?" 
    To: 'Søren Schmidt' <sos@DeepCore.dk>
Date: Mon, 30 Aug 2004 19:16:37 +0200

    Søren Schmidt wrote:

    > Hmm, looks like you re triggering the throtteling code, that
    > will lead to catastophic failure as it tosses out interrupts
    > causing this:

    Yes, that is very likely to be the source of my problems. However, the root
    cause is still not clear to me. Why would I get an interrupt storm when
    geom_stripe tastes the discs when everything else works like a charm?

    I have hw.intr_storm_threshold=5000 set in /boot/loader.conf so the storm is
    probably pretty "big". I added that setting back in June when, after
    installing a second RocketRAID 454 card, I started getting interrupt storms
    at boot during device probing. It didn't really help, but ACPI changes at
    the beginning of August seems to finally have removed those problems.

    > Anyhow you would want up to date -current ATA sources as quite a few
    > problems has been solved..

    Hmm, anything in the last 4 days that could affect this (kernel/userland was
    from 4 days ago)? I don't recall seeing any commits that would be relevant.

    /Daniel Eriksson

    _______________________________________________
    freebsd-current@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-current
    To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"

  • Next message: Maxim Sobolev: "burncd(8) usability: why `-s max' isn't default?"

    Relevant Pages

    • Re: sio: lots of silo overflows on Asus K8V with Moxa Smartio C104H/PCI
      ... Keep trying to make it use a fast interrupt. ... Another reson to try using a fast interrupt is that interrupt storm ... > input character rate. ... RELENG_4 has the correct formatting "clk irq0". ...
      (freebsd-current)
    • Re: 6-STABLE locks solid - current ok, why?
      ... managment, supend and resume support. ... Most system hangs are a result of lost interrupts or an interrupt storm. ...
      (freebsd-stable)
    • Re: 2.6.24-rc8 hangs at mfgpt-timer
      ... interrupt storm of some sort as soon as you register the interrupt handler. ... in 0x51400022 are the MFGPT interrupts ...
      (Linux-Kernel)
    • Re: [FreeBSD] Kann nicht drucken.
      ... Jeder versuch im Web-Frontend von CUPS eine Testseite zu drucken wird ... Im Handbuch war ich von der Stiefmüttzerlichen behandlung von CUPS ... Und was ist so ein interrupt storm? ...
      (de.comp.os.unix.bsd)
    • Re: Potential source of interrupt aliasing
      ... Doing this should result in an instant interrupt storm since your PCI ... and the machine will spend all it's time rescheduling the ithread. ... only have one CPU running? ...
      (freebsd-current)