Oracle has come underneath hearth for failing to repair a recognized problem with Home windows cases on its cloud infrastructure (OCI).
One consumer instructed The Register that regardless of the issue with boot failures on Home windows hosts inflicting manufacturing outages, Oracle has solely provided a workaround relatively than a repair.
Oracle published the workaround earlier this month. It mentioned that after restarting a Home windows compute occasion, the system fails in addition and stays caught on the loading display. It mentioned workaround choices included performing a diagnostic reboot, rebuilding the occasion, and restarting it.
Nevertheless, a disgruntled techniques administrator talking to The Reg mentioned the workaround was nonetheless inflicting manufacturing outages in workloads supporting SaaS functions.
“About six weeks in the past, we had a bunch of servers simply fail in addition. It was post-patching – Home windows safety – and we restarted mechanically, and so they simply did not boot. We tried the whole lot to get them again, and ended up restoring them from backup and creating new cases,” he mentioned.
Oracle initially prompt the problem might have been attributable to modifications on the consumer’s facet, relatively than examine. Nevertheless, Huge Pink later added the issue to its recognized points.
“Should you host an enterprise software that occurs to nonetheless use legacy VMs on Oracle Cloud, and people are operating Home windows… in case you restart your server, you are in a little bit of a lottery as as to whether it should come again up or not,” the sysadmin mentioned.
The tech professional defined that out of 60 manufacturing Home windows servers in OCI, 4 didn’t boot. Two of these had been a cluster, taking out one of many firm’s cases of the applying.
“It requires handbook interventions,” he mentioned. “The client has to go diagnose and restart, however there’s nothing that mechanically does this. From our standpoint, we do not have confidence whether or not a server goes to restart or not. We even have to use safety patches, as a result of that is simply good observe, and safety patches must restart, so it is inflicting manufacturing outages.”
The consumer, who requested to not be recognized, mentioned that the corporate principally used Azure, however saved some workloads on OCI as a result of they use Oracle databases. He mentioned if the reboot drawback persevered, the corporate would contemplate placing the OCI workloads in Azure, which could value extra in storag, however provides comparatively higher assurance of reliability.
El Reg has contacted Oracle for remark.
Oracle might resolve to rectify the issue as a sizzling repair – or out-of-cycle repair – however may be delayed due to communication with Microsoft, mentioned Iain Saunderson, CTO of Spinnaker Assist, which gives assist for Oracle and SAP.
“It is most likely one thing distinctive in Oracle’s setting that possibly Microsoft would not expertise wherever else. It does spotlight the funding you have to make and the belief you have to give to your cloud supplier to do their greatest relatively than simply doing what works for them.
“Oracle might be attempting to repair it, however they’re like, ‘lets have [a] work round and see if that buys us a cross on having to supply a sizzling repair or having to escalate this with Microsoft.’ It speaks to the belief that you just put in your cloud suppliers.” ®
Source link