Updated 07 Oct 2015
I’m sure you must have noticed the poor performance since the so-called upgrade when the site was moved from the Ramjet to the Thundercracker server six days ago.
When I submitted a support ticket I received this reply:
Performance will take some time to settle down as the new platform uses a multi-tiered SAN – however the SAN needs to tier data according to it’s usage before it performs optimally.
We’re sorry for caused inconveniences.
Server performance will be stabilized soon.
Please hold on.
The next few days showed a small improvement – until today. Then tonight I was advised:
Thundercracker and Snarl are scheduled for RAM upgrades to alleviate performance issues. This will require a reboot which will happen at 8pm tonight (05/10/2015).
It took a while to realise that the server was now back in service and access to the website was being stopped by a different factor. After examining the message “Warning: Unknown: failed to open stream: Permission denied in Unknown on line 0 Fatal error: Unknown: Failed opening required ‘/home/bigdavec/public_html/index.php’ (include_path=’.:/opt/php53/lib/php’) in Unknown on line 0” I decided to look at the index.php file that had been mentioned. It had been unchanged since October 25th 2013, but then I noticed that the file permissions had been set to 0000 which meant that it could not be accessed by anyone. After resetting to the usual value 0f 0644, suddenly the site became available again. I am trying to find out why those values were changed as I certainly didn’t do it.
06 Oct 2015
The following was added to the Evohosting status website today:
Hosting Platform Performance: 11:10 06/10/2015
As customers will be aware, we have recently performed upgrades to the existing Evo Hosting platform onto enterprise class hardware and storage solutions.
Extensive planning and measures were incorporated into the upgrades to minimise any disruption and as a result, the experience has been largely seamless for the majority of customers. However, we have received reports from some customers where they have experienced speed or performance issues since the upgrades.
Please be assured that our engineers and system administrators are aware of this and are currently addressing any issues as a matter of urgency.
Our new platform is more than capable of handling the hosting and email infrastructure, so the main reasons for any performance issues are largely as the new hosts process any of the imported data to assign if the correct tiers and priority levels. This has taken longer than we initially predicted.
To further expedite this process, we are currently adding further resource at a hardware level into the hosting and email cluster to boost overall resources, allowing performance to increase further still. At the same time, our senior network architects are also manually optimising the hosts to ensure they are working more efficiently.
We apologise for any disruption or inconvenience to customers who have experienced speed or performance issues. This is a temporary issue and will be resolved as soon as possible. Once complete, all customers will experience better performance and resilience than before, as part of the overall upgrade and enhancement plan.
We will continue to make updates to this status page with any further time-frames or updates.
Update: 16:35 06/10/2015
Our engineers have completed preliminary updates to the hosting platforms and performance has been increased for many customers.
We are still in the progress of updating all the hardware based performance upgrades, with the remaining tasks scheduled in for tomorrow morning
We are still making performance updates manually in the meantime.
Many customers will have seen speed increases, however if you are still experiencing delays or other performance problems, then these should largly be resolved when our updates are complete.
We will update this page again in the morning following confirmation from our Data Center engineers that the rest of the hardware based updates are live
Update: 13:05 07/10/2015
From approximately 2pm today, there will be brief unavailability to services as we further increase physical memory and other resources to our platforms.
We apologise for any disruption caused, however this will address customers facing performance issues, in particular database driven sites.