Blair Zajac wrote:
>I just discovered that I have a hung svn server (rev 4259).
>
>I've attached the server-status page. The server responds to the
>top level HTML page, but any requests to the repository hang.
>
>This is on RedHat 8.0 with db 4.1.25 and apr/apr-util and http2
>-r APACHE_2_0_BRANCH from 1/4/2003.
>
>Here's a backtrace.
>
># gdb -p 18403
>GNU gdb Red Hat Linux (5.2.1-4)
>...
>Attaching to process 18403
>Reading symbols from /opt/i386-linux/installed/apache-2.0-cvs-2003010400/sbin/httpd2...done.
>...
>0x420d3b2e in select () from /lib/i686/libc.so.6
>(gdb) bt
>#0 0x420d3b2e in select () from /lib/i686/libc.so.6
>#1 0x40215480 in db_xa_switch () from /opt/i386-linux/db-4.1/lib/libdb-4.1.so
>
This symbol is always wrong. It must be inlined or something, I believe
the correct symbol is __os_sleep() for #1 in the backtrace.
>#2 0x401eb2f1 in __os_yield () from /opt/i386-linux/db-4.1/lib/libdb-4.1.so
>#3 0x401663ef in __db_tas_mutex_lock ()
> from /opt/i386-linux/db-4.1/lib/libdb-4.1.so
>#4 0x401bb27e in __db_e_attach () from /opt/i386-linux/db-4.1/lib/libdb-4.1.so
>#5 0x401b76ce in __dbenv_open () from /opt/i386-linux/db-4.1/lib/libdb-4.1.so
>#6 0x4037576e in svn_fs_open_berkeley (fs=0xbffff1e8,
> path=0xfffffdfe <Address 0xfffffdfe out of bounds>)
> at subversion/libsvn_fs/fs.c:562
>#7 0x403626f2 in get_repos (repos_p=0xfffffdfe,
> path=0x811d5b0 "/export/home2/svn-ow/repos", locktype=1, open_fs=1,
> pool=0x8142d98) at subversion/libsvn_repos/repos.c:846
>#8 0x4036284f in svn_repos_open (repos_p=0xfffffdfe,
> path=0xfffffdfe <Address 0xfffffdfe out of bounds>, pool=0xfffffdfe)
> at subversion/libsvn_repos/repos.c:901
>#9 0x4034cd1d in dav_svn_get_resource (r=0x8130bc8,
> root_path=0x811d560 "/repos",
> label=0xfffffdfe <Address 0xfffffdfe out of bounds>, use_checked_in=-514,
> resource=0xfffffdfe) at subversion/mod_dav_svn/repos.c:1160
>#10 0x40313b0a in dav_get_resource (r=0x8130bc8, label_allowed=1,
> use_checked_in=-514, res_p=0xbffff33c) at mod_dav.c:705
>#11 0x40313d89 in dav_method_get (r=0x8130bc8) at mod_dav.c:790
>---Type <return> to continue, or q <return> to quit---
>#12 0x08066b65 in ap_run_handler (r=0x8130bc8) at config.c:195
>#13 0x0806717f in ap_invoke_handler (r=0x8130bc8) at config.c:401
>#14 0x08063996 in ap_process_request (r=0x40324288) at http_request.c:288
>#15 0x0805efcc in ap_process_http_connection (c=0x8130bc8) at http_core.c:293
>#16 0x08070235 in ap_run_process_connection (c=0x8124ab0) at connection.c:85
>#17 0x080651b8 in child_main (child_num_arg=-514) at prefork.c:696
>#18 0x08065314 in make_child (s=0x0, slot=46) at prefork.c:790
>#19 0x080655b9 in perform_idle_server_maintenance (p=0x809d750)
> at prefork.c:925
>#20 0x08065c50 in ap_mpm_run (_pconf=0x1, plog=0x80c77f8, s=0x1)
> at prefork.c:1120
>#21 0x0806bdcf in main (argc=3, argv=0xbffff694) at main.c:651
>#22 0x420158d4 in __libc_start_main () from /lib/i686/libc.so.6
>
>
>Note in
>
>#6 0x4037576e in svn_fs_open_berkeley (fs=0xbffff1e8,
> path=0xfffffdfe <Address 0xfffffdfe out of bounds>)
> at subversion/libsvn_fs/fs.c:562
>
>the odd address. Is this normal?
>
>
>
>
This is the exact same backtrace I get everytime mod_dav_svn dies.
Note, that I usually see other threads stuck in select() from other
spots as well.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jan 13 21:19:00 2003