#121 afpd SEGVs when accessing share with nonexistent dbpath

9.1.0.1
open
nobody
None
1
2013-07-19
2013-07-19
Alan Burlison
No

I have a share set up with an alternate location for dbpath. The share path is /mnt/emily and the dbpath is set to /mnt/.apf/emily. /mnt/.apf exists but /mnt/.apf/emily doesn't as the user in question has not yet logged in and attached to the share. Another user (dan) who has read-only access on the share has tried to access it, resulting in the following:

Jul 19 14:03:00 nasty cnid_metad[2079]: set_dbdir: mkdir failed for /mnt/.apf/emily
Jul 19 14:03:21 nasty last message repeated 22 times
Jul 19 14:03:21 nasty afpd[3234]: transmit: Request to dbd daemon (db_dir /mnt/emily) timed out.
Jul 19 14:03:21 nasty kernel: pid 3234 (afpd), uid 1003: exited on signal 6
Jul 19 14:03:21 nasty afpd[3234]: ===============================================================
Jul 19 14:03:21 nasty afpd[3234]: INTERNAL ERROR: Signal 11 in pid 3234 (2.2.4)
Jul 19 14:03:21 nasty afpd[3234]: ===============================================================
Jul 19 14:03:22 nasty afpd[22079]: AFP3.3 Login by dan
Jul 19 14:03:22 nasty afpd[22079]: afp_disconnect: trying primary reconnect
Jul 19 14:03:22 nasty afpd[2076]: Reconnect: no child[3234]
Jul 19 14:03:27 nasty afpd[22079]: afp_disconnect: primary reconnect failed
Jul 19 14:03:27 nasty cnid_metad[2079]: set_dbdir: mkdir failed for /mnt/.apf/emily
Jul 19 14:03:48 nasty last message repeated 22 times
Jul 19 14:03:48 nasty afpd[22079]: transmit: Request to dbd daemon (db_dir /mnt/emily) timed out.
Jul 19 14:03:48 nasty afpd[22079]: Reopen volume /mnt/emily using in memory temporary CNID DB.

Discussion