We have encountered stability issues: a storage failure that led to Exalogic being fully down' sometimes the storage cluster OCFS2 used for the OVM repository had stopped working and could not be mounted; finally, after many months, Oracle support discovered problems with InfiniBand (IB) host channel adapters (HCAs) that connect ZFS storage to the hosts and many firmware upgrades.

Sometimes while restarting the VM servers, they get stuck or no NIC is attached to the VM.

We faced, a few times, problems with the stability of the platform due to storage issues.

OTD (Oracle Traffic Director) is failover architecture and is not as robust as advertised. Additionally, SSO and several other features sometimes went down.

The stability is very good. Once you configure it, it runs with no issues. Out of ten I'd give it an eight.

We have had issues with the stability as the product failed on its promise for replication of all the components. There was a zpool failure which was exceptionally problematic.

I had an issue with the GFS of the storage.

Not anymore. Several months ago, we had some stability issues but, following patches delivered by Oracle, the machines are very stable.

Initially, yes, there were stability issues. A number of redesigns were needed to stabilize the Hypervisor, control the stack layer, as well as the software defined network drivers (Mellanox bugs causing compute nodes to reboot). The latter was addressed by a heavy work initiative, but required a set of patching, impacting customers' system stability.

We've had no issues with the performance.

Stability of Exalogic's been great. In a virtualized environment, runs OVM, doing a lot of work with it. In fact, we've been using some of the Exalogic here at this event to run some of our hands-on labs that I've been doing as community efforts and haven't had any issues with stability on it, no big panics. With any system, you do have to patch it. If you don't patch it for a year or two, you might start getting some bugs.

The root filesystem on the physical computer nodes began to fill-up as images were added with each quarterly PSU upgrade. We might want to expand that filesystem to account for that. Also, provisioning more than a dozen VMs can be problematic for the proxy servers to handle, and some of the jobs fail. There seems to be no issue with less than 10 VMs.

Stability for the most part has been good with some production impact which resulted from patching and version control. The product needs a better support model.

No stability issues.

The solution is stable, but it typically takes time to build up to stability. Sometimes there are some bugs. Overall, we find it to be quite reliable.

To be honest, there was a stability issue. As I mentioned, we had to live with some bugs for months.

Very stable.