[9fans] venti sync fixed? - Plan9

This is a discussion on [9fans] venti sync fixed? - Plan9 ; Hi, I installed a Plan 9 from a rather recent cd image (about two weeks old). With fossil+venti. Everything seemed to work well, however, after the second boot, I got the same venti sync problems as others have reported here. ...

+ Reply to Thread
Results 1 to 5 of 5

Thread: [9fans] venti sync fixed?

  1. [9fans] venti sync fixed?

    Hi,

    I installed a Plan 9 from a rather recent cd image (about two weeks old).
    With fossil+venti. Everything seemed to work well, however, after the second
    boot, I got the same venti sync problems as others have reported here.

    Now. On IRC it was rumoured that this issue was found and perhaps fixed. Is
    it? Could there have been a way for me to find out myself if this has been
    fixed? Is there a CHANGELOG for kernel/fossil/venti?
    --vk


  2. Re: [9fans] venti sync fixed?

    > Hi,
    >
    > I installed a Plan 9 from a rather recent cd image (about two weeks old).
    > With fossil+venti. Everything seemed to work well, however, after the second
    > boot, I got the same venti sync problems as others have reported here.
    >
    > Now. On IRC it was rumoured that this issue was found and perhaps fixed. Is
    > it? Could there have been a way for me to find out myself if this has been
    > fixed? Is there a CHANGELOG for kernel/fossil/venti?
    > --vk


    unless by "recent" you mean yesterday, you should try
    again. there has been a lot of change in the last month.

    - erik


    --rw-rw-r-- M 10912 geoff sys 2592 Sep 8 23:03 ifile.c
    --rw-rw-r-- M 10912 geoff sys 40524 Sep 17 20:52 fixarenas.c
    --rw-rw-r-- M 10912 geoff sys 4500 Oct 1 01:41 part.c
    --rw-rw-r-- M 10912 geoff sys 5659 Oct 1 12:48 config.c
    --rw-rw-r-- M 10912 geoff sys 14072 Oct 1 12:48 conv.c
    --rw-rw-r-- M 10912 geoff sys 10802 Oct 1 12:48 mirrorarenas.c
    --rw-rw-r-- M 10912 geoff sys 5776 Oct 1 12:48 verifyarena.c
    --rw-rw-r-- M 10912 geoff sys 4054 Oct 1 12:48 stats.c
    --rw-rw-r-- M 10912 geoff sys 16482 Oct 1 12:48 hdisk.c
    --rw-rw-r-- M 10912 geoff sys 4947 Oct 1 12:48 lump.c
    --rw-rw-r-- M 10912 geoff sys 5316 Oct 1 12:48 clump.c
    --rw-rw-r-- M 10912 geoff sys 20137 Oct 1 12:48 arena.c
    --rw-rw-r-- M 10912 geoff sys 15583 Oct 1 12:48 dcache.c
    --rw-rw-r-- M 10912 geoff sys 19110 Oct 1 12:48 dat.h
    --rw-rw-r-- M 10912 geoff sys 17933 Oct 1 12:48 index.c
    --rw-rw-r-- M 10912 geoff sys 10676 Oct 1 12:48 icache.c
    --rw-rw-r-- M 10912 geoff sys 22869 Oct 1 12:48 httpd.c
    --rw-rw-r-- M 10912 geoff sys 2310 Oct 1 12:48 checkarenas.c
    --rw-rw-r-- M 10912 geoff sys 4816 Oct 1 12:48 syncarena.c
    --rw-rw-r-- M 10912 geoff sys 21645 Oct 1 12:48 buildindex.c
    --rw-rw-r-- M 10912 geoff sys 5584 Oct 1 12:48 venti.c
    --rw-rw-r-- M 10912 geoff sys 9208 Oct 1 12:48 fns.h
    --rw-rw-r-- M 10912 geoff sys 1277 Oct 1 12:48 syncindex.c
    --rw-rw-r-- M 10912 geoff sys 1951 Oct 1 12:48 syncindex0.c
    --rw-rw-r-- M 10912 geoff sys 7528 Oct 1 12:48 icachewrite.c


  3. Re: [9fans] venti sync fixed?

    Vesa Kaihlavirta wrote:

    >Hi,
    >
    >I installed a Plan 9 from a rather recent cd image (about two weeks old).
    >With fossil+venti. Everything seemed to work well, however, after the second
    >boot, I got the same venti sync problems as others have reported here.
    >
    >Now. On IRC it was rumoured that this issue was found and perhaps fixed. Is
    >it? Could there have been a way for me to find out myself if this has been
    >fixed? Is there a CHANGELOG for kernel/fossil/venti?
    >--vk
    >
    >

    Here was no rumor on IRC. Russ just fixed it:
    http://plan9.bell-labs.com/sources/p.../nventi/readme
    Copy files from sources and rebuild it you want to test.

    cinap.



  4. Re: [9fans] venti sync fixed?

    On 10/2/07, Kernel Panic wrote
    >
    > >Now. On IRC it was rumoured that this issue was found and perhaps fixed.

    > Is
    > >it? Could there have been a way for me to find out myself if this has

    > been
    > >fixed? Is there a CHANGELOG for kernel/fossil/venti?
    > >--vk
    > >
    > >

    > Here was no rumor on IRC. Russ just fixed it:
    > http://plan9.bell-labs.com/sources/p.../nventi/readme
    > Copy files from sources and rebuild it you want to test.



    Ah, gotcha. I just tried with a very recent image, and had the same
    unfortunate behaviour.

    Any guesses how soon this fix will end up in the install CD?

    --vk


  5. Re: [9fans] venti sync fixed?

    > >I installed a Plan 9 from a rather recent cd image (about two weeks old).
    > >With fossil+venti. Everything seemed to work well, however, after the second
    > >boot, I got the same venti sync problems as others have reported here.
    > >
    > >Now. On IRC it was rumoured that this issue was found and perhaps fixed. Is
    > >it? Could there have been a way for me to find out myself if this has been
    > >fixed? Is there a CHANGELOG for kernel/fossil/venti?
    > >--vk
    > >
    > >

    > Here was no rumor on IRC. Russ just fixed it:
    > http://plan9.bell-labs.com/sources/p.../nventi/readme
    > Copy files from sources and rebuild it you want to test.


    Today's ISO does not seem to fix the problem on a qemu install.
    This is what I did:

    fetch the iso
    verify (md5 387cb470f2c46e5f8116ef7419086a05)
    bzcat < plan9.iso.bz2 > plan9.iso
    qemu-img create -f qcow2 img 10G
    qemu -monitor stdio -vnc :2 -cdrom plan9.iso -boot d img

    Then a usual install choosing fossil+venti.
    On install completion restart qemu.

    qemu -monitor stdio -net tap -net nic -vnc :2 img

    Login as glenda.

    kill timesync|rc # as it thinks it is Jan 2, 2000
    ndb/cs -f /lib/ndb/local
    ip/ipconfig
    ndb/dns -r
    aux/timesync -n # back to the present!

    cat >> /rc/bin/termrc.local
    TIMESYNCARGS=(-n pool.ntp.org)
    ndb/cs -f /lib/ndb/local
    ip/ipconfig
    ndb/dns -r
    ^D
    fshalt
    #wait for it to finish
    ^t^tr

    On reboot system pauses at ...init...sync... After a few
    minutes verify there is no disk activity due to qemu on the
    host system.

    If it matters this is on a qemu 0.9.0 (for i386), without
    kqemu, run on a 64 bit mode freebsd on amd64.

+ Reply to Thread