Since a few days, our redirectors are rejecting some clients with “Too many attempts to gain dfs [read|write] access to the file”. Are there any obvious causes to check when getting this error? Could this be shadowing an underlying cause that we should investigate?
The source code [0] indicates this is a client setting, but it appears that neither redirector/server (v4.10.0) nor client (v4.10.0) were changed when the error started occurring. We see the error for repeated SAM tests [1, 2] of which many are still passing, so the same setup on both sides *sometimes* fails and *sometimes* succeeds. Could there be some race condition we should take a look at?
[1] Redirector Log
/var/log/xrootd/manager/cmsd.log:191015 09:08:30 30012 SelDFS seeking /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/cmsd.log:191015 09:08:36 14961 SelNode
f01-120-181-e.gridka.de assigned /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/cmsd.log:191015 09:08:44 14961 SelNode
f01-124-182-e.gridka.de assigned /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:31 24030 Decode f01-031-107-e delays monalisa.1782:
[log in to unmask] 5 /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:36 24030 Decode f01-031-107-e gave monalisa.1782:
[log in to unmask] err -2 'Unable to access file; file does not exist.' /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:36 8693 monalisa.1782:
[log in to unmask] ofs_open: 100200-40644 fn=/13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:44 8876 monalisa.1782:
[log in to unmask] ofs_open: 100200-40644 fn=/13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:52 8693 monalisa.1782:
[log in to unmask] ofs_open: 100200-40644 fn=/13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
/var/log/xrootd/manager/xrootd.log:191015 09:08:52 24030 Decode f01-031-107-e gave monalisa.1782:
[log in to unmask] err -2 'Too many attempts to gain dfs write access to the file' /13/40572/97b83210-ef1a-11e9-bf12-0242a3ceac7a
cms.delay drop 45 qdn 3 servers 2 service 30 startup 30
cms.sched affinity none cpu 30 runq 50 io 10 mem 10 space 0 fuzz 100
cms.ping 20 log 1 usage 15
cms.space min 16g 16g
cms.dfs limit 0 lookup distrib mdhold 0 redirect verify retries 2