We have some good news for PHP users on the (gs) Grid-Service. Available for immediate use, (mt) Media Temple has released a new, simple method for dealing with PHP versions and updates. Please read on for more information.

Outdated software leaves you open to hackers, and updating old versions by hand is a pain. To keep you safe and to make site management simpler, we are now providing new version controls with automatic updates.

Updated Controls

The latest PHP versions are now available in the AccountCenter. For each domain hosted with Media Temple, you can choose which version to use.

We recommend choosing the most recent version, as it will be the most secure and ensure compatibility with the latest web apps. For instance, the latest version of WordPress requires PHP 5.2.4 or newer. If you’re using WordPress, be sure to upgrade! It’s pretty easy.

For those of you specifying PHP versions manually, you can learn more about the AddHandlers in our KnowledgeBase article.

Automatic Updates
As new PHP versions are available and others become unsupported, we will be removing old releases and automatically moving you to stable and secure versions. While most websites will be unaffected by these upgrades, we will make sure you receive plenty of notice so only good surprises come your way.

If you have any questions, feel free to add a comment here or join us in the User Forums.

Thanks for reading!

At the beginning of the year we mentioned several initiatives we’ve been working on to improve overall (gs) performance. Last month we told you about email performance improvements. Today we’re happy to tell you about ongoing storage improvements we’ve made that have improved overall web response times across the Grid.

For the last couple months we’ve been quietly enhancing the (gs) storage infrastructure with the addition of a caching layer built with extremely high speed Solid State drives. We’re about 80% complete with this work and happy to tell you that Grid Clusters with this new enhancement are reporting 25% speed increase for average web response.

As you can see above, storage segments using SSDs are responding faster. The graph shown here compares web response times from three (gs) storage segments with SSDs to three without. Before installation, trial storage performed roughly the same as other segments. As SSDs were added, response times decreased and settled around 25% faster than the rest.

This is a notable improvement, but just one enhancement in ongoing improvements we’re continuing to implement across the (gs) storage architecture.

What is a Solid State Drive?

Put simply, it’s a storage drive that has no moving parts. No spinning disks mean significantly faster data access times, though they generally can’t hold as much data as traditional disk. The speediness of SSD’s make them a perfect fit for a caching layer, where speed is of the utmost importance.

Why is Caching Important?

Good caching tends to make a big difference in a large, clustered environments like the (gs), because the most ‘popular’ data is what tends to slow down the traditional disks leading to increased response times. With a fast cache layer in place, the popular data comes directly from cache, taking the load off the rest of the storage system which leads to overall speed increases for all users.

Does this apply to my site yet?

Right now all customers on clusters 3, 4, 7 ,8 and 9 are using SSDs and work has started on clusters 1 and 6, with 2 and 5 scheduled immediately after. Upgrades are scheduled for upcoming normal maintenance windows to make the transition as convenient as possible. To find out which cluster your (gs) resides on, log in to the AccountCenter and click on your Server Guide from your grid control panel.

Ongoing Improvements

We’re happy with the results from SSDs and the Dovecot email rollout , but we’re not stopping there. We’ve recently started increasing the physical RAM across all (gs) nodes, and we’ll be reporting on that as it gets closer to completion.

Thanks for reading! Feel free to leave any comments below and join the discussion in the User Forums.

It’s time for some more news about the GRID. In this session, we’ll cover:

• Gen 2 Upgrade – 100% complete!
• MySQL stability
• More processing power
• Email improvements
• Cluster 6 — Coming Soon
• Updated 1-Clicks
• (cs) update

All Clusters Now Gen 2 Architecture!

You read right, the Gen 2 upgrade is 100% complete. As mentioned in the last update post, our final task was getting the backup system architecture on Clusters 1 and 2 into Gen 2 compliance. Now, should any unforeseeable frontline Storage Segment problems happen, we can quickly failover to the backup servers and minimize customer downtime. With this task complete, we now have architectural uniformity across all of our Clusters!

Performance Enhancements and Bug Fixes

With the Gen 2 Upgrade complete, we can focus on improving other aspects of your hosting experience. We are continuing to address known quality-of-service issues and enhancing the Grid’s performance characteristics.

MySQL Stability

We’re aware of the intermittent SmartPool connectivity issues some users are still experiencing and working on several concurrent projects that should ultimately lead to rock-solid MySQL reliability. Firstly, we are making some modifications to MySQL SmartPool server configurations for better performance. In addition, we are making enhancements to the Tahiti load-leveling engine that migrates high-load database users to isolated containers. Upon completion the system will be able to identify and migrate problem users faster than ever before, eliminating the “bad neighbor” effect that caused some of the connectivity issues our SmartPool users have experienced.

More Processing Power

The GRID is growing! To keep pace with organic growth, we’ve added new GRID nodes to almost all of our Clusters. This will yield better performance for your various services and provide a better experience for your viewers. In addition to that, we’ve been busy replacing older hardware with newer, faster machines.

Email Improvements

Email usage on mature GRID clusters continues to increase beyond projections, prompting some additional enhancements. Recent issues with email performance on Clusters 2 and 4 has led us to evaluate and improve our email response monitoring system and provide better quota tracking to reduce overall system load.

Other Grid News

Brand-new Cluster 6

After much backbreaking labor in our server rooms, we’ve powered everything on and the new cluster is being constructed. This should be done within the next few weeks to keep up with the demand of new (gs) customers.

Updated 1-Clicks

1-Clicks for Zencart and WordPress have been updated to their latest versions. If you have not already done so, please update these applications. 1-Click WordPress installs can be updated within the AccountCenter or from within your WordPress control panel.

(cs) Cluster-Server Update

Some time ago we announced our plans for the GRID’s successor, (cs) Cluster-Server. Those plans haven’t changed, but our number one priority remains making the (gs) a rock-solid, across-the-board stable product. As we continue to make improvements to the (gs) platform’s reliability, we’ll be able to renew focus on all the platform additions and enhancements that are required for (cs).

That’s all for now. Thanks for reading! Remember to join our community in the (mt) User Forums for more discussions and information!

It’s time for some more news about the GRID. In this session, we’ll cover:

• Gen 2 Upgrade – 99% complete!
• More Horsepower
• MySQL SmartPool
• Anti-Spam Update – RBL Reputations Continue Improving.

Gen 2 Upgrade Progress

We’re within 1% of a full Gen 2 Upgrade for Clusters 1 and 2! As stated in previous posts, this entire process was to be achieved as a two-stage project — with storage first and full-system architecture to follow. Here’s what we’ve already accomplished:

• Gen 2 storage transition (as mentioned previously) – Complete
• OS Redundancy and Gen 2 architecture — Complete
• Front-line Gen 2 architecture (web, email, databases)— Complete

The only thing that’s left to do is the final Gen 2 Backup Architecture Upgrade. Our engineers expect this will be achieved within two weeks.

Backup Architecture Upgrades

We’re working to duplicate the backup architecture that already exists on Clusters 3, 4 and 5 for C1/C2. This configuration allows us to immediately set a backup server to be a front-line data source, eliminating the need to copy the data from backups to a storage segment before being able to serve up the content.

More GRID News!

Front-line Servers Get Additional Horsepower

We are adding more front-line servers to the GRID. With these extra boxes will come the benefit of faster service response times and improved anti-spam reputations with third-party RBLs. Deployment is already in-progress.

MySQL SmartPool v.2 Update

In our last update we talked about our ongoing efforts to improve the SmartPool experience for underperforming segments. A solution to this issue is still in development, but we’ve completed a good portion of the work in recent weeks. We are on schedule for the remaining work that needs to be done and hope to have some good news in the next (gs) Update.

As we’ve said before, we have a special program for anyone who’s having an unsatisfactory experience with MySQL SmartPool v.2 while we work this all out. Contact our Support Team for more details.

Anti-Spam Update – RBL Reputations Continue to Improve.

Previously mentioned enhancements are paying off and we continue to see spam volume decrease. As a result, our status with most RBLs are in the green for all Clusters.

That’s all for now. Thanks for reading! Remember to join our community in the (mt) User Forums for more discussions and information!

It’s time for some more news about the GRID. In this session, we’ll cover:

• Gen 2 Upgrade – 95% complete
• Gen 1 Storage is in the dumpster!
• More GRID News.
• Anti-Spam Update – Better stopgaps. RBL reputations continue to improve.
• Recent System Incidents

Gen 2 Upgrade Progress

Looks like we’ve reached about 95% completion of a full Gen 2 transition for Clusters 1 and 2! Remember, this transition is happening in two phases: 1st Storage; 2nd Full-system architecture, and we are very near the end of the entire upgrade.

The priority of our engineers, now, is focusing on removing single points of failure by bringing online our Gen 2 OS Redundancy, followed immediately by upgrading the backup architecture to provide quicker recovery and failover to match Clusters 3, 4 and 5. Full OS Redundancy should be finished within a week with backup architecture to follow .

GEN 1 storage is in the dumpster!

The transition to Gen 2 storage on Clusters 1 & 2 is now finished! We have successfully migrated all customer sites to our Gen 2 storage systems powered by Sun — a milestone for (gs) Grid-Server. This means past incidents caused by Gen 1 storage will not reoccur in the future. More importantly, we can now begin decommissioning Gen 1 storage, thus unifying the storage platform across the entire GRID.

More GRID News!

Performance enhancements

We’ve added a sizable amount of storage performance to our clusters in past months. Customers are definitely noticing the increase in speed and reliability. Thank you for your feedback. We never stop making refinements to the GRID.

Better Storage Segment quotas

A recent System Incident revealed that no system is perfect and the unexpected can result in a difficult situation. As we outlined in our Incident Review of #867, we’re changing the formula for allocating sites onto our Gen 2 Storage segments. We are confident these changes will give us an improved safety net to catch issues before you do.

Sun Solaris 10 U7

The latest version of Sun Solaris 10 has been deployed to one of our forward-facing (gs) Storage segments (Cluster 2, Segment 07) and several of our backup segments. This OS release includes several enhancements that we are already seeing benefits from.

MySQL SmartPool Issues

System monitoring and forum discussions have shown that some of our customers have been experiencing database related problems. We have started working on a resolution for this we think our customers are going to enjoy. If you have any issues with your databases please contact our Support Department. We can provide you with a temporary solution.

Anti-Spam Update

We have been busy adding further stopgaps for better detection of outgoing spam from the GRID.

If we suspect an email you’re sending is spam, you’ll get a friendly notification pointing you to help resources. The spammers get shut out, and you get a helping hand. We have reduced ALL of our outgoing mail by 20%, accounting for a large chunk of spam that was previously leaving our network, and getting the GRID into trouble. With this significant reduction our reputation with RBLs has been improving, and will only continue to get better.

This new system still needs further refinement, but so far it has proven to be very effective from initial observations. We will continue refining our systems until we are satisfied that you are satisfied!

Recent System Incidents

We recently had a couple of System Incidents that we want you to know about.

1. Incident #867: Email access and web latency on Cluster 02
2. Incident #864: Brief Availability Issues on Cluster 05

Both of these incidents occurred on Clusters containing Gen 2 technology. It is important to note that the root causes for both of these incidents were not directly related to the platform itself. Here’s what went wrong:

• Incident 867 related to a rare instance of file-system fragmentation we had not previously encountered. This was related to the rapid movement of data off of the Gen 1 storage. For further details please see the incident review linked above.
• Incident 864 was related to a bad networking interface that has since been fixed. We have built corrective measures into our design process so that this segment has more redundancy features.

Thank you for reading. Remember, we’re always here to listen. Leave us a comment or join our (mt) User Forums to participate in our community discussions.

Your latest installment of timely (mt) information is here. This is what we’ll be covering in this week’s (gs) Update:

• Gen 2 Upgrade Status – 50% complete
• Fiji Agent
• Cluster 5 is alive!
• MySQL 5 BETA, Anti-Spam, DNS tool update

Gen 2 Transition Progress

We’re happy to announce that we are about 50% complete in the total upgrade process for Clusters 1 and 2! We continue to make steady progress in our transition efforts, recently adding new technology and hardware that allows migrations to move much quicker — without affecting overall Grid account performance.

As previously mentioned, this transition is a 2-phase project. First step: storage transition —migrating all Cluster 1 and 2 users on to Gen 2 storage devices. Next step: full system architecture updates built to Gen 2 spec (web, database, email and backup nodes).

Note: “Gen 2 Transition” only applies to those customers on Clusters 1 & 2 of our (gs) system. All cluster 3,4 and 5 customers are already on the Gen 2 system.

Gen 2 Storage Transition Update

Thanks to system enhancements, we’ve increased the rate of migrations significantly. And, since the last Grid update, we’ve moved roughly 3500 more customers to Gen 2 Storage. This gets us to about 75% completion in the first phase of the Cluster upgrade process —all done while maintaining full grid system performance.

Why is it taking so long?
Moving terabytes of data takes time. Couple that with the need to keep all customers’ sites up and accessible during the transition, and it makes sense that it’s a lengthy process. Because overall system performance is a top priority, we’re making the migration as quickly and safely as possible. That said, due to the amount of data left to move, we aren’t quite ready to give a solid estimate on completion.

Total Gen 2 Upgrade Status

As mentioned above, total upgrade progress (storage and architecture) is at about 50%. Before a total Gen 2 upgrade can be completed, we first need to complete phase 1 by moving the remaining 25% of sites over to Gen 2 storage. As we get closer to completing the storage transition, we’ll provide detailed information and timelines on the final phase of the Gen 2 upgrade.

Fiji Agent Update

As stated in previous posts, we have developed a tool for those Cluster 1 and 2 customers who would like to take advantage of the Gen 2 System without waiting for the Cluster Upgrade process to be completed. If you think you’d like to speed up your move to the new technology, perhaps Fiji Agent is the answer. Please submit a support request for more details or to make a migration request.

For detailed information about 1st Generation vs. 2nd Generation and Moving Old Customers to New Technology, read our post Anatomy of Storage on the GRID.

General (gs) Updates

That’s all for this post. Thanks for taking the time to read through it. Check back regularly for news and information on all things (mt). Or, perhaps, you may want to dive deeper. In that case, join the (mt) User Forums and give us some feedback.

Still need more? Get real-time updates by following us on Twitter.

In this week’s (gs) Update, we cover the ongoing progress of transitioning (gs) Clusters 1 and 2 to Gen 2 architecture, more news about Cluster 5, an update on our new Fiji and MySQL5 tech, and a bite-sized Anatomy Of piece explaining the finer details of our transition to Gen 2 storage.

Cluster 5 Progress

We are nearly done with getting Cluster 5 online, with an expected arrival within the next week. As we work to upgrade both Gen 1 clusters, we must also sustain our growth. Equal priority is being placed on both objectives. Our recent maintenance on May 10 was one of the last hurdles remaining to bring Cluster 5 online.

Storage Transition

As you may know, a portion of our customers on Cluster 2 recently experienced an extended service interruption. This was covered in detail in System Incident #754. Recovery was a multi-phase process, with our Engineers working around the clock to restore data. While resolving the incident itself was a time-consuming process, we made a lot of smart moves that ultimately helped expedite the overall recovery time.

In keeping with “best practices” we always maintain an inventory of spare equipment in the event of emergencies. In this particular incident we were able to use Gen 2 hardware that was already reserved for our planned transition. Essentially we were able to lay the groundwork necessary for our complete transition away from our older architecture. We currently have around 15,000 sites remaining on Gen 1 hardware. Approximately 2,300 sites were moved to a new storage segment based on Gen 2 hardware as part of resolving the incident.

Gen 2 Hardware vs Gen 2 Architecture

For a cluster to be considered Gen 2 it requires our newer Sun-based hardware, in addition to other dramatic architecture changes as displayed below:

While we are converting all remaining storage segments to Gen 2 hardware, we are essentially in a Gen 1.5 environment for these two clusters. This is an in-between phase until we complete the transition. Over the next few months we will finalize this change and report our progress. Since this will require significant work, most of these changes will occur during off-peak hours in the form of Scheduled Maintenance actions.

Fiji Agent

We have been actively processing cluster move requests with our new Fiji Agent technology. Our engineers have scaled this technology dramatically in the past weeks, allowing our Support Department to move even more sites in parallel. We can even tap into this same power to migrate sites between storage segments on the same cluster, something we were able to do in the past but can now do much more quickly. This technology will prove vital in the coming weeks.

As noted in previous posts, Fiji will ultimately be our migration engine designed to work across all our product lines and datacenter locations. It will be a vital pipeline within the (mt) Media Temple product family.

MySQL 5 BETA for Clusters 1-3

In our last update we mentioned that we were almost ready with MySQL 5 for our first 3 Clusters. A couple of weeks ago we added our initial batch of testers and squashed a few big bugs right away (THANK YOU to those who helped with this first round of testing!). We have another batch ready to go this week. If you are interested in following our progress, we encourage you to join the discussion threads in the (gs) MySQL 5 BETA section of our (mt) User Forums.

We wanted to let you know about the progress we have been making with the (gs) Grid-Service platform. Coming soon we’re going to devote a section of our weblog to keep you up-to-date on what’s been keeping us busy. Have a look at what our (gs) team has been doing.

INFRASTRUCTURE ENHANCEMENTS

06/06/2008 – Git libraries added:
A version control system that is all the rage right now. Search our KnowledgeBase to Git with the program!

06/10/2008 – RubyGems Updated:
The latest versions and a much easier installation process. Get your gems all at once now!

05/27/2008 – PHP updated to 5.2.6:
The latest stable PHP release that includes all of the security patches is now available for our shared hosting platform.

05/19/2008 – Email load drastically reduced:
A more robust authentication system was brought online to get rid of those pesky errors that have plagued us in the past. The “421; too much  load, try again later” errors have become a thing of the past.

05/19/2008 – @mail upgrade to 5.2:
The @mail webmail software used on the (gs) has been upgraded to 5.2. We also plan to release more frequent updates to this software in the future as they become available.

05/13/2008 – DNS gets a facelift:
PowerDNS launched.  Response times have improved from around the globe for all our hosted services.  Internal processes, like zone file changes, are quicker now too.

05/06/2008 – Storage keeps on getting better:
We’ve learned some things and made some changes. The result: fewer Scheduled Maintenances, the end of the “micro-outage”, stability improvements, and a re-architected method of attaching our storage arrays to clusters; this means happier times for everyone!

04/10/2008 – Subversion updated:
We bumped the version of SVN up to 1.4.2 on the Grid. We’re using the latest version officially supported by the Debian Operating System.

03/03/2008 – Improved SSH sessions:
Your connections should be more stable. In addition, we have increased our security checks to prevent DDOS attacks.

–Ongoing – MySQL:

SmartPool v2 improved:
Since launching SmartPool v2, we’ve taken our experience and combined it with consultations with the MySQL team at the 2008 MySQL Conference & Expo. We’ve made significant improvements to the SmartPool v2 system. A basic one is lower density SmartPool servers while doubling the amount of available pool servers. This has shown to improve baseline performance for pool users and brings us closer to eliminating the “bad neighbor” effect!

–Other MySQL Improvements

11/01/2007 – Tweak your MySQL easily and instantly:
Users of our (gc) MySQL GridContainer technology now have the ability to make changes to their my.cnf directly from the WebControl! This is big news for anyone who wants their database server set up just the way they like it. This new feature also includes the ability to “roll back” to the default configuration!

10/09/2007 – Multiple database users for SmartPool v2:
You asked and we delivered! Based on customer feedback and popular demand from our User Forums, we upped the limit on MySQL users for SmartPool v2 sites to 5. You can have up to 5 users without needing a container, and you have granular control over each user’s permissions.

ACCOUNTCENTER

03/31/2008 – 1-Click Magento:
“Open source e-commerce evolved.” Now just a click away!

03/31/2008 – 1-Click Upgrader for WordPress2.5:
This is just the beginning.  We will implement this with our other installers soon!

03/11/2008 – uControl:
With this new feature we helped make everyone’s life easier! If you’re the “website guy/gal” for your company, you no longer need to be the one changing everyone’s passwords and email preferences; they can now do it all on their own, freeing you up to get more important stuff done…

COMING SOON

(mt) Labs

Summer ’08 – Django:
We have gotten tremendous amounts of feedback and positive refinements from our (gc) Django GridContainer BETA program. Expect a product launch coming very soon including a couple of unexpected surprises!

Fall ’08 – MySQL 5:
MySQL 5 is coming to the (gs)!. We’re currently getting things right for the beta and will be contacting all of our applicants with more info soon!

Winter ’08 – Cluster-Server:
The successor to the (gs) Grid-Service. Stay tuned to the (mt) Labs page for updates, new information, and a future private BETA test!

On the Horizon – MailProtect gets tougher:
Spam is in for a bumpy ride through our pipes. Using the latest bleeding-edge tools we will take email protection to the next level!

On the Horizon – Backup Tool Restored and more:
We’re waving goodbye to the File Manager as you know it.  Get ready for more functionality and operations, including the backup/restore features you’ve come to expect from us!

Our engineering team has upgraded @Mail to version 5.0.4.  This webmail upgrade for the Grid comes with many exciting new features outlined below.

New Features:

• faster performance
• enhanced email searching
• improved spell check support
• improved address book importing & management
• better html email viewing & editing
• hundreds of minor bug fixes
• now powered by PHP

All Grid-Service customers may immediately begin using this upgraded version by visiting yourdomain.com/webmail, webmail.yourdomain.com or by clicking through the AccountCenter. If you’ve been using the previous version of @Mail many clients have reported the need to clear their browsers cache for an optimal experience. We hope you enjoy the new enhancements, and look forward to your feedback!