Hello. Recently we've installed new version of SRSS - 4.1 on our SXCE b97 x86-64 Solaris server. It worked great on test stands, we have modified gvfsd and hal - now connected usbdisks are recognized as hal devices shown on the desktop... But in production environment I have serious problems. There are a lot of terminals with 26D SRSS errors. The reason for this is zombie dtlogin sessions. Randomly session is recognized as zombie session and dtlogin stops monitoring its Xservers(Xnewt).
I have the following info in dtlogin logs: (main) StopDisplay(): :4, server pid = 10927, manager pid = -1, shutdown = 1 (main) Resetting keyboard (main) WaitForSomething (main) Processing child death, pid = 10927 (main) Account: id=dt0A, user=NULL, line=NULL, pid=10927, type=DEAD_PROCESS (main) PamAccounting: prog_name=dtlogin (main) PamAccounting: display_name=:4 (main) PamAccounting: entry_id=d t 0 A (main) PamAccounting: user=NULL (main) PamAccounting: line=NULL (main) PamAccounting: line_dev=/dev/NULL (main) PamAccounting: pid=10927 (main) PamAccounting: entry_type=8 (main) PamAccounting: exitcode=0 (main) PamInit: LC_MESSAGES=C (main) PamInit: LANG=ru_RU.UTF-8 (main) PamInit: NLSPATH=/usr/dt/lib/nls/msg/%L/%N.cat:/usr/lib/locale/%L/LC_MESSAGES/%N.cat:/usr/dt/lib/nls/msg/C/%N.cat:/usr/lib/locale/C/LC_MESSAGES/%N.cat: (main) PamInit: (DISPLAY exported) display_env=DISPLAY=:4 (main) PamInit: pam_start success (main) PamInit: pam_set_item returned 0 for item 3001, value (2.0) (main) PamInit: display_name = :4 (main) PamAccounting: entry_type is DEAD_PROCESS (main) PamAccounting: Closing open PAM Handle for DEAD_PROCESS (main) Enter: solaris_reset_utmp_mgmt() (main) reset_utmp:ut_pid 11020, pid 10927 (main) PamAccounting: DEAD_PROCESS reset_utmp error=27 (main) Zombie server reaped, removing display :4 (main) Removing display :4 from display list. -- This message posted from opensolaris.org
