Installation guide

Acopia Networks, Inc.
(“longnameddirectory”) actually reside on the other share’s directory (with the actual name of
“LONGNA~1”). Clients cannot access the files unless someone runs another sync files operation.
(28748)
• A ron evict command can result in a growing packet storm among the RON’s remaining ARXes,
possibly resulting in one or more ARX reboots. (28889)
• If a volume’s CIFS client attempts to set the FILE_ATTRIBUTE_OFFLINE attribute (often used in
stubs for filer-ILM applications), and one of the volume’s filers cannot accept it, the volume may lock
out all client access.
Current Solution: Specific errors in the syslog indicate that this attribute was rejected by a particular
filer, and F5/Acopia personnel have tools to work around the issue. Contact F5 Support if a CIFS
volume blocks out client access and you see any TRANFS-…-INVOFFLINE or TRANFS-…-
INVPARAM errors in the syslog. (28829)
• A no share, remove-share, or similar share-removal operation can be excessively slow in a large
shadow volume, or in any managed volume with thousands of link files. (28494)
• A shadow volume may copy more files than were selected by the source fileset. (29021)
• If a file or directory in a source volume is renamed between shadow copies, and the new name is
different from the old name in case only (e.g., “myfile.txt” becomes “MYFILE.txt”), the later shadow
copy fails with a STATUS_OBJECT_NAME_COLLISION error. This only occurs in a shadow
volume that supports CIFS. (28809)
• MTU path discovery is not negotiated properly. (28807)
• Configuring the same XIPs on both switches can prevent a VIP from starting. (28804)
• During DNS lookup, the domain name is translated to lowercase regardless of what is supplied by the
user or configured on the switch, and the lookup fails. (28808)
In Release 2.7.0
• A memory fault in the Physical Address Extension (PAE) can lead to unpredictable issues, including
an nsck destage that never completes, new VIPs that are unreachable, or even a chassis reboot. (27455)
• The slot erase command does not properly clean up the configuration database. (27756)
• Shadow-volume statistics restart from 0 (zero) after a shadow-copy run is interrupted by a failure (such
as a reboot). (27046)
• The initial walk of a shadow-copy rule runs to completion after encountering a common error (such as
a file opened by a CIFS client), then it restarts from the beginning. (27651)
• Volume software can fail (and write a core file) when a CIFS client has files open in two different
volumes and the client’s connection is lost. (27678)
• Certain client applications may crash unless the CIFS path-cache feature is disabled. (27013)
• NFS clients for a presentation volume experience occasional timeouts when accessing storage that is
attached to a managed volume in the same VPU. (27934)
9