Hi folks. I have been doing large numbers of parallel releases of volumes
of various size, but they are all in the 60 - 120G range. This is happening
over a WAN.
Volume A and A.readonly are on primary RW fileserver
Volume A.readonly replica is on another fileserver, on the other side of a
100Mbit WAN link.
It seems to take a very long time to release (which is known.. I am not too
concerned about that), but I am suddenly coming into this problem:
audio-2013.236
RWrite: 536878929 ROnly: 536878930 RClone: 536878930
number of sites -> 3
server array3.foo.cc partition /vicepz RW Site -- New release
server array4.foo.cc partition /vicepc RO Site -- Old release
server array3.foo.cc partition /vicepz RO Site -- New release
This is a completion of a previous release
Starting transaction on cloned volume 536878930... done
Deleting extant RO_DONTUSE site on array4.foo.cc... done
Creating new volume 536878930 on replication site array4.foo.cc: done
Starting ForwardMulti from 536878930 to 536878930 on array4.foo.cc (full
release).
Release failed: rxk: sealed data inconsistent
Could not end transaction on a ro volume: rxk: authentication expired
Could not update VLDB entry for volume 536878929
Failed to end transaction on the release clone 536878930
Could not release lock on the VLDB entry for volume 536878929
rxk: authentication expired
Error in vos release command.
rxk: authentication expired
I realize I can delete and add back the replica, and retry the release,
however I am concerned about the error. How is it that an authentication
can expire when the client is automatically reinitiated and aklog'd every 6
hours?
--
Timothy Balcer / IT Services
Telmate / San Francisco, CA
Direct / (415) 300-4313
Customer Service / (800) 205-5510
