This is the one everybody has been begging for! croit v1901 with high available SMB and iSCSI gateways and other incredible features that will make your already amazing cluster even better. Although it’s already been available since January, we only now got around to writing this blog post, so here are some more details on the update.
FEATURES AND CHANGES:
Ceph 13.2.4
croit v1901 comes with a new server image for Ceph 13.2.4. The most important feature in Ceph 13.2.4 is the new cache size auto-tuning: you can now set a target memory usage per OSD process instead of explicitly configuring the cache size. The auto-tuner will then adjust the cache size to match the total memory usage, i.e., it will temporarily reduce the cache size if other tasks such as recovery require more memory temporarily. We set the memory target to 3 GB per OSD by default, you can adjust it in the Ceph config view.
Highly available SMB gateways
The new image also contains a new version of Samba with a libcephfs backend. You can now configure SMB shares through the UI, just add them like any other service to a set of servers. SMB comes with its own high-availability management for seamless failovers, you can configure the virtual IPs to be used by the servers directly in the SMB configuration. All configured IPs will be distributed evenly across all healthy SMB gateways.
This is the first version of our SMB gateway and it only features two static users for now. We plan to ship full Active Directory integration in the next release.
HIGHLY AVAILABLE ISCSI GATEWAYS
The new image also comes with an iSCSI gateway based on ceph-iscsi that can be configured in the UI. It supports multi-pathing and fail-over.
ceph-iscsi currently has a few restrictions that we can’t circumvent at the moment: the most important one is that you cannot remove a gateway once you add it. So be careful when adding servers to a gateway.
S3 EXPLORER
You can now browse objects in your S3 buckets directly from the GUI. It’s also possible to configure ACL settings, upload files, and download files.
TASK ADVISOR
Our flexible and powerful configuration systems allows you to change a lot of low level settings. But with great power comes great responsibility: it is possible to misconfigure clusters and that’s something we want to avoid.
The new task advisor module catches suboptimal settings and advises you on how to fix it. For example, you will see it telling you to upgrade the server image to the latest version after upgrading croit to v1901.
LAST BUT NOT LEAST
- Our main menu got a little bit cluttered over the past years and we’ve cleaned it up in this release
- Statistics view loads faster now
- Better support for LDAP nested groups in active directory environments
- Duplicate IP checks: croit will now warn you if you configure the same IP address on two different servers
- Autocomplete and help for custom ceph.conf options
- CephFS snapshots in the file explorer
- Optimized upload of multiple small files
UPGRADE NOTES: INCOMPATIBLE CHANGES
We completely removed support for legacy kernel-based NFS exports. Please migrate to the new NFS server before upgrading, old NFS exports will no longer be configured after rebooting a server. Old kernel-based NFS exports currently show up as “legacy NFS gateway” and are marked with a warning icon. Any NFS gateway created in the past year is not affected.
We moved time-series data from Python to Go code to speed up the statistics view. This migration also changed how external monitoring systems like Grafana access the stats: statistics are now served from the same endpoint as the other API. Change your monitoring URL to https://{croit-mgmt}/stats/ Create a croit user with Viewer permissions for your monitoring system.
This change only affects you if you use our embedded time series database; users who configured an external Graphite service are not affected.
