Professional Web Applications Themes

streams going haywire - SCO

Removed by Administrator...

  1. Moderated Post

    Default streams going haywire

    Removed by Administrator
    Stuart J. Browne Guest
    Moderated Post

  2. Moderated Post

    Default Re: streams going haywire

    Removed by Administrator
    Stuart J. Browne Guest
    Moderated Post

  3. #3

    Default Re: streams going haywire

    Stuart J. Browne wrote:
    > > I have a box of which has the streams buffers going haywire (see class 7
    > > streams allocations below).
    > Is there any way I can track down what's chewing these resources so I can
    > track down and fix this issue?
    >
    > The only thing that's changed on the system recently is we went away from
    > VisionFS and started using samba, but it isn't currently being used to
    > serve files, just printing. It was installed from the 2.2.6 volumes on
    > Skunkware.
    >
    > Anybody have any ideas how I can track down this one?
    google groups "samba getsockopt streams leak patch"... Three
    particularly apropos articles (each can be a standalone solution):

    [url]http://groups.google.com/groups?selm=20030426004547.J27257sco.com[/url]
    [grab ip/{Driver.o,space.c} from 507]

    [url]http://groups.google.com/groups?selm=bXKdndHuQ4xAIyajXTWcrgcomcast.com[/url]
    [Brian K. White has built prepatched binaries]

    http://groups.google.com/groups?selm=b76fkd$12ma$1FreeBSD.csie.NCTU.edu.tw
    [source patch, and it's been checked into ongoing SAMBA 2.8 source]
    >Bela<
    Bela Lubkin Guest

  4. #4

    Default Re: streams going haywire


    "Bela Lubkin" <belalsco.com> wrote in message
    news:20030628062446.GM22421sco.com...
    > Stuart J. Browne wrote:
    >
    > > > I have a box of which has the streams buffers going haywire (see
    class 7
    > > > streams allocations below).
    >
    > > Is there any way I can track down what's chewing these resources so I
    can
    > > track down and fix this issue?
    > >
    > > The only thing that's changed on the system recently is we went away
    from
    > > VisionFS and started using samba, but it isn't currently being used to
    > > serve files, just printing. It was installed from the 2.2.6 volumes on
    > > Skunkware.
    > >
    > > Anybody have any ideas how I can track down this one?
    >
    > google groups "samba getsockopt streams leak patch"... Three
    > particularly apropos articles (each can be a standalone solution):
    >
    > [url]http://groups.google.com/groups?selm=20030426004547.J27257sco.com[/url]
    > [grab ip/{Driver.o,space.c} from 507]
    >
    > [url]http://groups.google.com/groups?selm=bXKdndHuQ4xAIyajXTWcrgcomcast.com[/url]
    > [Brian K. White has built prepatched binaries]
    >
    >
    http://groups.google.com/groups?selm=b76fkd$12ma$1FreeBSD.csie.NCTU.edu.tw
    > [source patch, and it's been checked into ongoing SAMBA 2.8 source]
    >
    And I remember reading that thread a few weeks/months ago.. What got me
    was that the problem didn't go away after I stopped the SMBD processes.

    Will implement one of these.

    Thanks Bela.

    bkx


    Stuart J. Browne Guest

  5. #5

    Default Re: streams going haywire

    On Sat, 28 Jun 2003 18:22:49 +1000, Stuart J. Browne wrote:
    >
    > "Bela Lubkin" <belalsco.com> wrote in message
    > news:20030628062446.GM22421sco.com...
    >> Stuart J. Browne wrote:
    >>
    >> > > I have a box of which has the streams buffers going haywire (see
    > class 7
    >> > > streams allocations below).
    >>
    >> > Is there any way I can track down what's chewing these resources so I
    > can
    >> > track down and fix this issue?
    >> >
    >> > The only thing that's changed on the system recently is we went away
    > from
    >> > VisionFS and started using samba, but it isn't currently being used to
    >> > serve files, just printing. It was installed from the 2.2.6 volumes on
    >> > Skunkware.
    >> >
    >> > Anybody have any ideas how I can track down this one?
    >>
    >> google groups "samba getsockopt streams leak patch"... Three
    >> particularly apropos articles (each can be a standalone solution):
    >>
    >> [url]http://groups.google.com/groups?selm=20030426004547.J27257sco.com[/url]
    >> [grab ip/{Driver.o,space.c} from 507]
    >>
    >> [url]http://groups.google.com/groups?selm=bXKdndHuQ4xAIyajXTWcrgcomcast.com[/url]
    >> [Brian K. White has built prepatched binaries]
    >>
    >>
    > http://groups.google.com/groups?selm=b76fkd$12ma$1FreeBSD.csie.NCTU.edu.tw
    >> [source patch, and it's been checked into ongoing SAMBA 2.8 source]
    >>
    >
    > And I remember reading that thread a few weeks/months ago.. What got me
    > was that the problem didn't go away after I stopped the SMBD processes.
    >
    > Will implement one of these.
    >
    > Thanks Bela.
    >
    > bkx
    Perhaps there is some new windows virus or spyware or web server attack
    going around that is causing a spike in traffic not previously seen?

    I have a customer with a 5.0.5 box (running facetwin, not samba or
    visionfs). the box is visible from the internet and there are about 25 pc
    in-house. nothing much has changed in about a year. however I too, just
    yesterday got a call from them about a problem with a doent scanning
    mechanisn not working. Part of the system uses the web server to display
    the scanned images back to the user and it wasn't working. everything
    seemed to be running, the images were getting deposited in the appropriate
    directories, the web server was running, i stopped & started httpd anyways
    (the stock netscape server in this case), still no go, but a possible clue
    that something was amiss on the box at a general, lower level: several
    time while browsing around the box using mc to view logs and look for
    files image files, my connection just winked out with no warning.

    That's when I noticed a _lot_ of NSTRPAGES messages in syslog.

    In my case it was mostly class 10, a few in a few other classes but class
    10 was collosal. The box has be under exactly the same sort of load for
    quite a long time so I decided not to touch the kernel and just had them
    reboot, working on the assumption that what probably happened is some
    program went crazy rather than that their usage actually exceeded their
    settings.

    Everything works fine again after the reboot.

    it's been a day and a half now and netstat -m still shows that their
    maximum usage so far has not reach the half-way mark, and the fail column
    is all 0's and the current usage (a "dead" time right now) is only about
    1/3 of configured.

    I had never seen this before so it's interesting to me that your post came
    along just now.


    current netstat -m for this box after a day and a half:

    streams allocation:
    config alloc free total max fail
    stream 1280 84 1196 12857 222 0
    queues 908 171 737 25717 447 0
    mblks 2148 899 1249 149621516 2140 0
    buffer headers 2362 1301 1061 1466564 2311 0
    class 1, 64 bytes 256 7 249 70711582 251 0
    class 2, 128 bytes 64 32 32 16327234 62 0
    class 3, 256 bytes 288 264 24 29789917 282 0
    class 4, 512 bytes 40 28 12 44705 33 0
    class 5, 1024 bytes 68 2 66 318328 67 0
    class 6, 2048 bytes 158 128 30 9290597 157 0
    class 7, 4096 bytes 109 98 11 650 109 0
    class 8, 8192 bytes 4 0 4 694 4 0
    class 9, 16384 bytes 3 0 3 144 3 0
    class 10, 32768 bytes 3 0 3 780 3 0
    class 11, 65536 bytes 0 0 0 0 0 0
    class 12, 131072 bytes 0 0 0 0 0 0
    class 13, 262144 bytes 0 0 0 0 0 0
    class 14, 524288 bytes 0 0 0 0 0 0
    total configured streams memory: 2512.00KB
    streams memory in use: 792.12KB
    maximum streams memory used: 1232.41KB

    The only network or service related problems I know of are:

    1) there are two laser printers plugged directly into the server via
    parallel ports and shared using facetwin, and several people print to them
    both from windows and from the unix application. One of these printers
    broke a few weeks ago and one of the parallel cables has been unplugged
    and several print jobs have been aculating in the spooler for that
    printer. (I cleared them and disabled that printer when I saw this
    yesterday)

    2) there are 4 or so facetwin printers, that is, printers hung off of PC's
    that are configured into the unix spooler via facetwin. One of these has
    always been problematic in that the PC in question is a fairly cruddy old
    install of windows 98 or maybe even 95 I don't recall., and that pc is
    also quite some distance away physically in a non-to-delicate environment.
    dirty, hot, probably not very clean power etc... this printer fairly
    commonly becomes disabled in the spooler due to not being reachable when a
    print job is submitted. (once every couple weeks to a month)

    3) the network is all connected via two medium-old 10/100 hubs (not
    switches)


    I wasn't going to put much thought or worry into actually diagnosing this
    one because it just so happens that I am installing a brand new server for this
    customer in a couple days, and as such, the time for tuning the kernel is
    kinda past. :)

    This customer has sarcheck installed, generating a nice html report every
    night and dumped in a directory under htdocs and each night makes a new
    file so all files going back to install are available to track trends.
    .... except it was only a demo that was never purchased and made it's last
    informative report on 05/15/2002. DOH! *smack* :)

    --
    Brian K. White -- [email]brianaljex.com[/email] -- [url]http://www.aljex.com/bkw/[/url]
    +++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
    filePro BBx Linux SCO Prosper/FACTS AutoCAD #callahans Satriani

    Brian K. White Guest

Similar Threads

  1. Scheduled Tasks went haywire this morning
    By sonicsight in forum Coldfusion Server Administration
    Replies: 1
    Last Post: March 8th, 02:09 PM
  2. Scheduled Tasks Seems To Have Gone Haywire
    By MichaelSJudd in forum Coldfusion Server Administration
    Replies: 0
    Last Post: February 21st, 11:36 PM
  3. Trial Version gone haywire
    By Zaxterion in forum Macromedia Flash Player
    Replies: 1
    Last Post: December 4th, 06:08 PM
  4. Spacing gone haywire
    By kelly Jean in forum Macromedia Contribute Connection Administrtion
    Replies: 2
    Last Post: March 7th, 07:05 PM
  5. save action gone haywire?
    By Michael Appelmans in forum Adobe Photoshop 7, CS, CS2 & CS3
    Replies: 2
    Last Post: July 8th, 10:33 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139