On Monday 19 June 2006 09:01, Ph. Marek wrote:
> The client side (doing a commit) hangs with
> write(4, "change-file-prop ( 4:c470 9:svn:" ..., 4096 <unfinished ...>
> while the other side (svnserve, called by ssh) does
>
> 08:58:12.865719 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.878338 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.888718 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.898829 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.908707 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.921092 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.930662 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.940865 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.950751 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> 08:58:12.960672 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
> and so on, and so on ...
> Any help?
BTW, that seems reproducible.
I killed both sides and started afresh, but the same symptoms appear again.
Maybe that helps:
0xffffe410 in __kernel_vsyscall ()
(gdb) bt
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xa7de2d8d in select () from /lib/tls/i686/cmov/libc.so.6
#2 0xa7ca3a9e in __os_sleep () from /usr/lib/libdb-4.3.so
#3 0xa7ca3b8d in __os_yield () from /usr/lib/libdb-4.3.so
#4 0xa7bf5ef3 in __db_tas_mutex_lock () from /usr/lib/libdb-4.3.so
#5 0xa7c87b56 in __lock_get_internal () from /usr/lib/libdb-4.3.so
#6 0xa7c88891 in __lock_vec () from /usr/lib/libdb-4.3.so
#7 0xa7c66242 in __db_lget () from /usr/lib/libdb-4.3.so
#8 0xa7c07123 in __bam_search () from /usr/lib/libdb-4.3.so
#9 0xa7bfa273 in __bam_c_rget () from /usr/lib/libdb-4.3.so
#10 0xa7bfb49c in __bam_c_rget () from /usr/lib/libdb-4.3.so
#11 0xa7c5b6c3 in __db_c_get () from /usr/lib/libdb-4.3.so
#12 0xa7c61cf5 in __db_get () from /usr/lib/libdb-4.3.so
#13 0xa7c646c2 in __db_get_pp () from /usr/lib/libdb-4.3.so
#14 0xa7ce39e8 in svn_fs_bdb__get_txn () from /usr/lib/libsvn_fs_base-1.so.0
#15 0xa7cedf9f in svn_fs_base__youngest_rev () from /usr/lib/libsvn_fs_base-1.so.0
#16 0xa7cef278 in svn_fs_base__rev_get_txn_id () from /usr/lib/libsvn_fs_base-1.so.0
#17 0xa7cef47d in svn_fs_base__rev_get_root () from /usr/lib/libsvn_fs_base-1.so.0
#18 0xa7cef504 in svn_fs_base__rev_get_root () from /usr/lib/libsvn_fs_base-1.so.0
#19 0xa7cef87a in svn_fs_base__record_undo () from /usr/lib/libsvn_fs_base-1.so.0
#20 0xa7cefa48 in svn_fs_base__retry_txn () from /usr/lib/libsvn_fs_base-1.so.0
#21 0xa7cee99f in svn_fs_base__begin_txn () from /usr/lib/libsvn_fs_base-1.so.0
#22 0xa7ec815e in svn_fs_begin_txn2 () from /usr/lib/libsvn_fs-1.so.0
#23 0xa7ed65b3 in svn_repos_fs_begin_txn_for_commit () from /usr/lib/libsvn_repos-1.so.0
#24 0xa7ed25ed in svn_repos_get_commit_editor2 () from /usr/lib/libsvn_repos-1.so.0
#25 0xa7e9016a in svn_ra_svn__drive_editorp () from /usr/lib/libsvn_ra_svn-1.so.0
#26 0xa7e8ed64 in svn_ra_svn__drive_editorp () from /usr/lib/libsvn_ra_svn-1.so.0
#27 0xa7e8cf87 in svn_ra_svn_drive_editor () from /usr/lib/libsvn_ra_svn-1.so.0
#28 0x0804f5cd in ?? ()
So it seems to hang at some mutex.
But I don't see anything:
# ipcs
------ Shared Memory Segments --------
key shmid owner perms bytes nattch status
------ Semaphore Arrays --------
key semid owner perms nsems
------ Message Queues --------
key msqid owner perms used-bytes messages
I'm the only user on both machines (AFAIK :-), and the repository is only used by one thread.
fuser on the database files shows only one pid.
Any ideas?
Thank you!
Regards,
Phil
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jun 19 09:23:10 2006