Improved mbrestore script to handle backups from different methods
Issues: MB-3360
Bug in ketama hashing algorithm. Note that this will potentially change the key->server hashing for items in memcached-type buckets. It is advised not to run with either Membase server nodes or Moxi processes of mixed versions.
Issues: MB-3799
Moxi used to use "Out of Memory" for any error. This has been resolved.
Issues: MB-3479
Added button to "flush" memcached buckets
Issues: MB-2765
Rebalance failed to restart after stopping
Issues: MB-3255
Moxi will now ignore invalid nodes of a cluster that it is connected to
Issues: MB-3575
OSS Couldn't join cluster with hostname on Windows
Issues: MB-746
Enhancements to bucket deletion code
Erlang bug made the Membase UI break with certain timezones
Issues: MB-1872
rpm -U doesn't restore data files after upgrade to 1.7.0
Workaround: Need to re-create a symlink as per the 1.7.0 upgrade instructions
Issues: MB-3922
Membase fails to run on Windows system with PGP (http://www.pgp.com) installed
Workaround: Uninstall PGP on any systems running Membase
Issues: MB-3688
TAP doesn't handle "garbage" data/messages over a dump connection
Issues: MB-3669
Unsynchronized clocks can lead to UI not displaying graphs
Workaround: Make sure all the clocks (and their timezones) are synchronized in a Membase cluster. Best to use NTP or some other synchronization technology
Issues: MB-2833
Can't edit size of Memcached bucket type
Issues: MB-2343
Linux installer tries to start (and fails) "membase-server"
Workaround: Another benign warning
Issues: MB-3797
Memcached process is not shutdown quickly with a high disk write queue
Workaround: Make sure your disk write queue is as low as possible (or 0) before shutting down the service
Issues: MB-3628
Membase-Server does not start on windows 2003 after successful installation
Workaround: Need to install vc redistributable (8.0.59193) on the system
Issues: MB-4013
Windows upgrade failed from 1.6.0 to 1.7.0
Workaround: Workaround is to follow upgrade instructions from 1.6.0 to 1.6.5 and then 1.6.5 to 1.7.0
Issues: MB-3932
Multi-get can return multiple values during rebalance
Issues: MB-2220
Need better error reporting when trying to run binary as root.
Workaround: Do not run as the root user. Note that this is not the same as installing, which needs to be done by root.
Issues: MB-1475
Known issues when Membase runs out of disk space
Workaround: Monitor your system resources closely and ensure that the Membase nodes don't run out of disk space
Issues: MB-3158
Known issues with 'flush_all' against Membase-type buckets
Workaround: 'flush_all' is not currently supported for Membase buckets due to some known issues. It should not be used in a production environment and the "flush" button is not available via the GUI for this reason
Known issues with membase-server (re)starting
Workaround: When a process is killed or dies, mnesia files might be left in a poor state that prevents restart.
Issues: MB-4006
MB-3333 MBS Network Filesystems (CIFS/NFS) are not supported for data file storage. Block-based storage (iSCSI,EBS, etc) should be fine
Issues: MB-3333
mbrestore sometimes shows benign python interrupt error.
Workaround: This should only happen at the very end of the restore and does not affect the data being inserted
Issues: MB-3804
Ubuntu 8.x is unsupported
Issues: MB-2506
Some known issues when rebalancing with a very high ratio of data on disk versus data in RAM
Workaround: Workaround is to use more nodes than less to better distribute the on-disk data
memory utilization is not accounted for correctly in certain situations
Workaround: Workaround is to set the memory quota (either bucket or server level) lower than normal to prevent any swap usage
Can't add 1.7.0 node to 1.6.x cluster
Workaround: Rolling upgrades are not currently supported from 1.6.x to 1.7.0. Use the in-place upgrade instructions.
Issues: MB-3924
UDP access blocked by malformed request
Issues: MC-158
Amazon "micro" and "small" instances are not supported
Workaround: It is a best practice to use at least "large" instances to provide enough RAM for Membase to run
Issues: MB-3911
Large changes in wall-clock time negatively affect disk persistence ability and GUI display
Workaround: Use NTP to keep systems synchronized and start Membase service after such synchronization has taken place
TAP stats sometimes show up as negative
Issues: MB-3901
'ejectNode' is possibly dangerous to active servers
Workaround: Only use 'ejectNode' on nodes not currently responsible for any vbuckets
Issues: MB-3423