On Friday, June 24, 2011 (at approximately 14:06 MDT), some ImageSilo users may have received an error stating that “The disk is full” when attempting view documents. We do not believe that this affected all ImageSilo users that were actively using the system at the time. The issue was caused by an extremely abnormal usage pattern that caused the storage for the cache location (where documents are temporarily cached to provide secure access to documents without exposing the entire document store to the internet) to be quickly consumed and not released. Engineers were alerted by our internal monitoring systems. By14:22 MDT (18 minutes later), the issue was fully rectified.

ImageSilo engineers have taken steps to ensure this issue doesn’t reoccur in the future, including increasing the cache size exponentially. We sincerely appreciate your business and apologize for any inconvenience this may have cause those who experienced the error.

ImageSilo Administration

This Service Bulletin Update is in regards to the Emergency Maintenance that was declared earlier this afternoon (on March 24, 2011). After thorough investigation and discussion with the load balancing hardware vendor, ImageSilo engineers were able to rectify the issue. Engineers will continue to closely monitor the systems.

ImageSilo Administration

 This Service Bulletin Update is in regards to the Emergency Maintenance that was declared on March 3, 2011 and continued on March 7, 2011. Last night, ImageSilo engineers applied the suggested Microsoft hotfix to the database clusters in order to address the performance-related bug. They have been monitoring the performance of the systems since that time. Based on data gathered since the hotfix application, the performance issue appears to be resolved. Engineers will continue to closely monitor the systems.

ImageSilo Administration

This Service Bulletin Update is in regards to the Emergency Maintenance that was declared on March 3, 2011 and continued on March 7, 2011. ImageSilo engineers have been working continuously with Microsoft to identify the cause of the performance issues related to ImageSilo’s database clusters. After evaluating a significant amount of data, Microsoft has made two recommendations – (1) make a change to a setting to allow SQL Server’s cache plan to be optimized for ad hoc queries, and (2) apply a hotfix to address a bug related to the performance issues (the hotfix was released a couple weeks ago, but has not made it into standard updates).

The change to optimize the cache plan was made at approximately 11:50 MST. Immediately following the change, there was an expected further degradation of performance (while the cache plan was flushed and rebuilt), which lasted about 15 minutes. Afterwards, performance returned to normal and currently appears to remain good. Engineers will continue to monitor performance going forward.

This evening, at 20:00 MST, ImageSilo engineers will be applying the suggested hotfix to fix the associated performance-related bug. This will require failing over the database clusters. As such, two short periods of inaccessibility (up to 5 minutes each) are expected this evening during the failover.

Questions/Comments: Please feel free to contact ImageSilo Administration with any questions or comments.
866.374.3569 (303.493.6900)
siloadmin@imagesilo.com