Archived

This forum has been archived. Please start a new discussion on GitHub.

Log File Size error

xdm
xdm La Coruña, Spain
Hi

I have and IceBox service that produce a big log i forget to rotate the log and server crash when log rich the 2GB (i thinks this a limit of the filesystem not sure now), but Ice don't throws any error when try to restart the service. Simple crash with signal 25, I remove the log and all works fine again. But was dificult figure what the problem is. Maybe you can throw and especific expcetion in future relase.

thanks

Comments

  • benoit
    benoit Rennes, France
    Hi,

    On which platform does this happen? Could you get the stack trace of the process in order to see where the crash occurs?

    Thanks,

    Benoit.
  • xdm
    xdm La Coruña, Spain
    Hi benoit

    this is in gentoo linux machine, linux-2.6.18, gcc version 4.1.1
    (Gentoo 4.1.1-r1),glibc 2.4-r4
    Ice-3.1.1 c++ patched whith Ice oficials pachs.

    Filesystem is jfs and have enough free space when it crash,I have
    other files bigger than 2 GB with the same user that is used to run
    icebox services. (seems like this is not a filesystem limit).

    ulimit -a report this
    core file size          (blocks, -c) 0
    data seg size           (kbytes, -d) unlimited
    max nice                        (-e) 0
    file size               (blocks, -f) unlimited
    pending signals                 (-i) 8122
    max locked memory       (kbytes, -l) 32
    max memory size         (kbytes, -m) unlimited
    open files                      (-n) 1024
    pipe size            (512 bytes, -p) 8
    POSIX message queues     (bytes, -q) 819200
    max rt priority                 (-r) 0
    stack size              (kbytes, -s) 8192
    cpu time               (seconds, -t) unlimited
    max user processes              (-u) 8122
    virtual memory          (kbytes, -v) unlimited
    file locks                      (-x) unlimited
    
    I can try to reproduce this in a develop machine, and send you a stack trace.

    I'm not sure what is the best way for get the stack trace in this
    case, because crash ocurrs when i try to start the service from
    icegridadmin, must I attach gdb to icegridnode and the try to start
    the service?

    My Ice distribution is compiled with out debug info, must i compile in
    debug mode ?

    thanks
  • benoit
    benoit Rennes, France
    Hi,

    You should be able to have log files with a size greater than 2GB with this OS. I don't know why it doesn't work for your IceBox server and seeing the stack trace could help.

    You should enable core files with "ulimit -c unlimited". Next time your server crashes a core file will be created and you'll be able to get the stack strace with gdb (with "gdb myserver core.<pid>").

    The best would be to run it with debug information enabled but it should also work without debug information (but we won't see line numbers for example).

    To get the stack trace of the crash, you can also attach to the process with the debugger once it's activated or run manually the server within the debugger using the configuration file generated by the IceGrid node.

    Cheers,
    Benoit.