Rackspace Cloud Files is their older object storage solution based on OpenStack Swift.
Although their status page was completely green, for the last two weeks we've been hitting weird errors and disconnections when accessing their London cluster.
Since this Monday, the error rate has grown greatly and we found that, not only we had problem sometimes when writing to the object storage ( something we could try to mitigate retrying the upload ) but also some files that where confirmed as uploaded by the API, weren't later available to download and gave 404s.
After a lengthy call with their support this morning they've retroactively updated their status page, but the only response I have to the issue is: "Engineers have increased the storage capacity in the region; however, it may take a few days for mitigation to take effect." A few days!
The status update is hilarious:
Further additional storage capacity will continue to be added in a controlled manner balancing object replication and deletion process to further restore the environment's health. Engineers continue to recommend that customers retry upload and delete actions as needed as this is the only workaround currently available.
Did they forget that their storage was filling up?
I know that not even Rackspace knows what the hell that company is about, but letting their storage fill up is beyond ridiculous.
The impact in our business has been, lets say, noticeable. And of course, we're migrating away as fast as we can transfer everything we have there.
Sorry, didn't see your comment. We have had tickets unanswered for days, and they refused to raise the ticket level to a different level other than "Standard", even when we and our clients have been heavily impacted for two weeks.
We migrated from LON to IAD, but after a couple days we started getting write errors on that zone too, on peak US hours, so we finally made a big effort and migrated all the object storage to GCP (which is faster and at a fifth of the price).
Rackspace is quite a joke right now. They have lots of overlapping products, legacy offerings, they don't even know if they do consulting or cloud or what. And the "famous Fanatical Support"... let me laugh.
Although their status page was completely green, for the last two weeks we've been hitting weird errors and disconnections when accessing their London cluster.
Since this Monday, the error rate has grown greatly and we found that, not only we had problem sometimes when writing to the object storage ( something we could try to mitigate retrying the upload ) but also some files that where confirmed as uploaded by the API, weren't later available to download and gave 404s.
After a lengthy call with their support this morning they've retroactively updated their status page, but the only response I have to the issue is: "Engineers have increased the storage capacity in the region; however, it may take a few days for mitigation to take effect." A few days!
The status update is hilarious:
Did they forget that their storage was filling up?I know that not even Rackspace knows what the hell that company is about, but letting their storage fill up is beyond ridiculous.
The impact in our business has been, lets say, noticeable. And of course, we're migrating away as fast as we can transfer everything we have there.