This problem was first reported by Seppo Nuutinen on 29 Oct 1997. The symptom was:
mvfs: ERROR: view=snuutine_vu vob=/vob/cgi-bin - ClearCase vob error
This usually happens when you have NFS problems. This time this was not the case. Recently, we got also errors in the view log (on the view server):
######### /var/adm/atria/log/view_log ########## 04/12/00 17:48:53 view_server(4837): Error: Unexpected error in VOB "feanor:/vobstorage/tms-os/tools10.vbs" op vob_ob_get_gen_num:error detected by ClearCase subsystem See vobrpc_server_log on host feanor 04/12/00 17:48:53 view_server(4837): Warning: Unable to open vob 'feanor:/vobstorage/tms-os/tools10.vbs': unknown error in VOB ##########
Log files on the vob server showed that the problem was with lockmgr defaults:
######### /var/adm/atria/log/vobrpc_server_log ########## 10/29/97 09:48:36 vobrpc_server(7634): Error: Too many open databases on host (try increasing -f argument on lockmgr command line) 10/29/97 09:48:36 vobrpc_server(7634): Error: Cannot open database in "/lun1/vobs/nnmsc/dust.vbs/db" 10/29/97 09:48:36 vobrpc_server(7634): Error: Error from VOB database: "/lun1/vobs/nnmsc/dust.vbs". 10/29/97 09:48:36 vobrpc_server(7634): Error: Trouble opening VOB database: "/lun1/vobs/nnmsc/dust.vbs". ##########
For a quick help, umount some of the obviously not needed VOBs.
Next I edited /usr/atria/etc/atria_start on the vob server to raise "-f" limit to 512 (Note the FAQ at Rational):
------------
# ${ATRIA}/etc/lockmgr ${LOCKMGR_ALIAS_OPT} -q 1024 -u 256 -f 256 >>$ADM_DIR/log/lockmgr_log 2>&1
${ATRIA}/etc/lockmgr ${LOCKMGR_ALIAS_OPT} -q 1024 -u 256 -f 512 >>$ADM_DIR/log/lockmgr_log 2>&1
-----------
One can do a kill -3 to the lockmgr, which results in
a report in the lockmgr_log. Esp. check a line like the following (the
number of open files is 7 times the number of vobs in use; force use
e.g. by change directory to each root):
52 users, 273 open files