vodsaccourorp

Starting on Monday September 19th we will be migrating Mars shared hosting accounts over to a new shared hosting specific server dal006 with the hostname dal006.hawkhost.com. The reason for this is currently Mars is one of our mixed servers which over the course of 2011 we have been migrating mixed servers to strictly shared or reseller. The one major advantage is once migrated over shared hosting users will have access to the shared hosting specific features that are currently not available on our mixed servers. It also means no longer will reseller users be able to affect services with their user accounts. In this case it will also be a hardware upgrade moving from a Dual Xeon 5520 with 24gb ram and 4x300GB Raid-10 to a Dual Xeon X5670 with 24gb of memory and 4x600GB Raid-10.

Since all our IP addresses are routed to our VLAN's rather than a specific server we can make this migration very seamless. We will migrate data then re-route the IP address over to the new server. This dramatically reduces the chances of data sync issues which are usually associated with server moves due to IP addresses changing. That is not to say it cannot happen but we are available to assist with any issues users may run into if for whatever reason their data does go out of sync. As for any downtime associated with the migration we estimate it'll be about 1 minute at most as we need to remove the IP from the old server and route it to the new one which can take us about a minute.

How the migration will work is we will start migrating users on dedicated IP's first starting on Monday. After that is completed we will move onto users with shared IP addresses. We hope that this entire migration will only last two days we hope but we've created a window of three days.

The only thing to keep in mind is once migrated it is not advised to be connecting to the mars hostname if you currently are. While the data will stay there for now eventually all shared hosting user data will be removed from the Mars server as it will be fully converted to a reseller hosting only server. So make sure to connect to the dal006.hawkhost.com server once the window has been completed.

If you have any questions regarding the maintenance do not hesitate to contact us. Our system administration department is available to answer any questions regarding this.


Start Date: 09/19/2011
End Date: 09/21/2011
Estimated Down Time: less than 5 minutes
Duration: 3 days

On Saturday September 15th between 1pm EDT and 5pm EDT we will be upgrading all our servers PHP 5 to PHP 5.2.4. We do not expect any down time while we perform this upgrade we will just be performing an apache restart once PHP 5.2.4 is installed. If you're interested about any changes in this version please visit http://www.php.net/ChangeLog-5.php#5.2.4.


Date: 09/15/2007
Start time (EDT): 1:00pm
End time (EDT): 5:00pm
Duration: 5 hours (few minutes per server large window to perform maintenance across all servers)

On Friday July 17th we will be upgrading the MySQL version of the Mars server to 5.0.81 from 5.0.77 between 12:30AM and 1:00AM CDT. We do not estimate that it'll take 30 minutes to do we will just be doing it during this window. We estimate there will be 1 minute of MySQL down time in order to restart the service to complete the upgrade. There may be none at all we're just giving an estimate that it could be 1 minute.

The reason for this upgrade is to stay on top of the latest versions available to us through cPanel.


Date: 07/17/2009
Start time (CDT): 12:30am
End time (CDT): 1:00am
Duration: 30 Minutes
Estimated Down Time: 1 Minute

On Wednesday December 17th at 11:00AM EST we will be upgrading the Apache and PHP on the Mars server to the latest versions. We do not expect any interruption in services however the CPU usage may be higher than normal on the server while we compile the new versions. We expect this upgrade to take no longer than 30 minutes.

Date: 12/17/2008
Start time (EST): 11:00am
End time (EST): 11:30am
Estimated Down Time: None
Duration: 30 minutes

Earlier this morning Mars experienced a brief bout of slowness / crashed. After a reboot everything is online and optimal.

It appears the issue was something exhausted all of the memory in the server rapidly causing SWAP usage / high disk IO that ultimately brought the machine to a halt. We're currently investigating the origins.

On Wednesday July 11th starting at 2PM CDT we will be taking the Mars server down to run a file system check due to some reported errors after reviewing system logs. We estimate the actual file system check will take less than an hour, however we've left the window open for 3 hours in the event we run into unexpected issues.

Date: 07/11/2012
Start time (CDT): 2:00PM
End time (CDT): 5:00PM
Duration: 3 hours
Estimated Down Time: 60 minutes or less

Throughout this window our support team will be available for any questions. They can be contacted via ticket at https://support.hawkhost.com.

We will be updating the kernel on the Yoda VPS node to the latest version between 10:00pm and 2:00am EDT. The reason for this is for security reasons while we do use ksplice the age of the kernel and the time it takes to patch this system at this point makes sense to just update the kernel itself. Even if we had to patch the system the virtual machines on it would have had to be all taken off line in order to do the updates at this point. So we believe it would be wise to just do an actual kernel update.

The maintenance window is for 4 hours but we estimate the down time will only be 5 minutes for the node and between 20 and 30 minutes for all virtual machines to be up and running again.

Date: 09/21/2010 - 09/22/2010
Start time (EDT): 09/21/2010 10:00pm
End time (EDT): 09/21/2010 2:00am
Duration: 4 hours
Estimated Down Time: 5-30 minutes

March was a shortened month as we only started doing the new monitoring on the 14th. So keep in mind this dataset is inaccurate and does not reflect the overall uptime that month actually had. I'm going to try to explain any downtime reflected on the reports as well as we generally know why something went down. This report also includes a few images of the uptime. Feel free to use them as you wish.

We will be updating the kernel on the Yoda VPS node to the latest version between 1:00am and 3:00am EDT.

The maintenance window is for 2 hours however the estimated downtime is 0 to 10 minutes.

Date: 06/06/2011
Start time (EDT): 06/06/2011 1:00am
End time (EDT): 06/06/2011 3:00am
Duration: 2 hours
Estimated Down Time: 0-10 minutes

The Yoda server is currently off line we have no idea at this time what exactly happened. We're currently having datacenter technicians investigate on site as we have no theories as to why the machine has gone off line completely including our IPMI management system.

On Saturday between 10pm EST and 12am EST we will be migrating Mars, Mercury and Venus to new primary IP addresses. The reason for this is to expand out our web server VLAN in order to accommodate more machines. Once this is done it will now be possible to easily move secondary IP's from one server to another. We do not expect any down time during this time however anyone referencing the servers by their main IP can no longer do this.

Once this is completed over the next few months we have plans of upgrading Mars, Mercury and Venus to much higher specification machines gradually. Thanks to the maintenance on Saturday we will also be able to transfer accounts to the upgraded servers without the need for dns changes or any down time or issues what so ever.

As always if you have any questions about the maintenance please contact our support department.


Date: 7/14/2007
Start time (EDT): 10:00pm
End time (EDT): 12:00am
Duration: 5-10 minutes

On Thursday May 12th between 1:00pm and 11:00pm CDT we will be upgrading the Mars server's PHP 5.2.x to the latest 5.2.x version. We do not expect any down time during the maintenance window. We will be upgrading the system PHP to the latest version first then the web servers so for a short time there may be different versions handling cron jobs and web server requests. The reason for this is to get various fixes for LSAPI as well as general PHP fixes as well.

If you have any questions about this maintenance window do not hesitate to contact support.

Date: 05/12/2011
Start time (CDT): 1:00pm
End time (CDT): 11:00pm
Estimated Down Time: None
Duration: 11 hours

On Saturday February 20th we will be upgrading the Mars server to the latest version of cPanel 11.25. We anticipate this upgrade taking no longer than an hour however with us doing many machines we cannot give a concrete time when each server will be done.

We do not anticipate any service interruptions during this time. cPanel may be slower than usual when the upgrading is being performed though. cPanel 11.25 brings many great new features and has reached stable and has been tested by us in a production environment on a limited number of servers.

If you have any questions about this maintenance window do not hesitate to contact us.

Date: 02/20/2010
Duration: 24 hours
Estimated Down Time: None

Starting on Monday September 26th we will be migrating Mars server to new higher end hardware as well as better system configuration when being strictly a reseller only system. We feel this will benefit all users on the Mars server as it will make it consistent with other reseller only servers we have. We estimate this will be completed no later than Wednesday September 28th.

We expect no more than 5 minutes of downtime per reseller account as we will be moving accounts to the new server then re-routing the IP address to the new hardware. This means there will be no changes to DNS and as a result no dns progation issues. In our experience our migration method means the large majority of users will not even realize that they've switched servers. If you do however run into any issues do not hesitant to contact us as our system administrators are prepared to resolve any issues experienced.

The new Mars server will be similar to that of our recent deployment of servers:

Old Mars:
Dual Xeon E5520
24GB RAM
4x300GB Raid-10

New Mars:
Dual Xeon X5670
24GB RAM
4x600GB Raid-10

So this is a pretty big upgrade as it will have four additional CPU cores with 8 to the operating system counting hyper threading. It will also be using 600GB disks over 300GB disks of the old system.

Once the migration starts the current mars system will be named oldmars while the new one will become Mars this is one thing to keep in mind during the migration.

If you have any questions regarding the maintenance do not hesitate to contact us. Our system administration department is available to answer any questions regarding this.


Start Date: 09/26/2011
End Date: 09/28/2011
Estimated Down Time: less than 5 minutes
Duration: 3 days

We will be removing Zend Framework from the PHP default include path on Tuesday October 6th. It was a great at the time to have for developers not wishing to upload the framework. Unfortunately as more applications make use of it there are problems that are starting to come up. One such popular example is Magento does not function properly without modifications to the .htaccess file.

This does not mean you cannot use Zend Framework on our hosting. You can still do this it's just you need to upload it to your account. This is probably better anyways as each version brings changes. So anytime we were upgrading it could have potentially broke your application anyways.

If you have any questions or concerns about this please contact support.

One of the drives on the raid-10 array of Yoda has failed. This was causing large amounts of i/o wait on the node as the drive had not yet been dropped. Once the drive was dropped from the raid array due to inconsistencies the i/o wait has come back down.

We are already preparing to replace the bad drive as our machines support hot-swap which means we can replace the drive on the fly. We do not have a time as of yet as we're in the process of coordinating with datacenter staff.

We will update this thread once we have an exact time on when the bad drive is being replaced.


Date: 03/21/2008
Start time (EST): 10:00pm EST
End time (EST): Ongoing
Estimated Down Time: None
Duration: Unknown

April was unfortunately a tough month at Devoted Host as our primary datacenter SoftLayer was running on generators due to the inclement weather in the area. When switching back to utility power a 2500amp breaker failed placing all our servers on UPS backup power. Unfortunately after 30 minutes machines started to go offline. After about 3 hours of work from SoftLayer electricians onsite power was restored to servers.

I'd like to make things clear here things happen from time to time that are unexpected. Hardware does fail the best you can do is restore service as fast as possible. The 3 hour timeframe this was done in was significantly better than other datacenters in the past have replaced breakers of this size.

So here is our uptime reports for all our servers it includes the no longer active Apollo as well as the now active Jupiter.

This maintenance applies to all shared, reseller, VPS and backup servers

Scheduled Maintenance - GLOBAL [06/18/2011-06/19/2011]
Date: Saturday, June 18, 2011 (06/18/2011) and Sunday, June 19th, 2011 (06/19/2011)
Start Time: 01:00 AM CDT
End Time: 05:00 AM CDT
Services affected: Public / Private Network
Location: GLOBAL
Duration: 28 hours

Event summary:

Datacenter engineers will be performing global maintenance on their backbone. During this maintenance, engineers will be implementing a number of policy changes related to route announcement and accepting routes from providers and peers. This work is required to support continued growth within the network and future global expansion. Due to the scope of this change, the work has been scheduled across the following dates:

Saturday June 18th 01:00 - 05:00 CDT
Sunday June 19th 01:00 - 05:00 CDT

Customer impact:

While engineers expect no outage during this maintenance, customers may experience higher latency and/or packet loss while routes reconverge across the network.

On Monday July 2nd we will be performing the upgrade to cPanel 11 on all servers. We will be doing this throughout the day and do not expect any interruption of web server service. We do however expect periods of time where cPanel will be unavailable due to the upgrade. We have no exact estimate on how long this will take or an exact time each server will have the upgrade performed. If you have any questions about this do not hesitate to ask our support department.


Date: 5/21/2007
Start time (EDT): 10:00am
End time (EDT): 10:00pm
Duration: 12 hours (Estimated)

There will be emergency network maintenance on FCR01 which our Mars server is on. This window is on Monday November 30th starting at 12:30am CST and lasting until 1:30am CST. Although the window is for 1 hour we estimate that there will only be a down time of 15-20 minutes. This maintenance is necessary to correct an issue that could cause network interruptions that would not be scheduled.

Further information about this can be found at: Emergency Network Maintenance [11/30/2009]


Emergency Data Center Maintenance - FCR01.DAL01 [11/30/2009]
Date: Monday, November 30, 2009 (11/30/2009)
Start time (CST): 00:30
End time (CST): 01:30
Services affected: Public Network
Location: DAL01
Duration: 1 hour

This maintenance applies to all shared, reseller, VPS and backup servers

Scheduled Maintenance - GLOBAL [07/22/2011]
Date: Friday, July 22nd, 2011 [07/22/2011]
Start Time: 12:00 AM CDT
End Time: 04:00 AM CDT
Services affected: Public / Private Network
Location: GLOBAL
Duration: 4 hours

Event summary:

Datacenter engineers will be performing global maintenance on their backbone. During this maintenance, engineers will be implementing a number of policy changes related to route announcement and accepting routes from providers and peers. This work is required to support continued growth within the network and future global expansion.

Customer impact:

While engineers expect no outage during this maintenance, customers may experience higher latency and/or packet loss while routes reconverge across the network.

This maintenance applies to all shared, reseller, VPS and backup servers

Scheduled Maintenance - GLOBAL [06/25/2011-06/26/2011]
Date: Saturday, June 25, 2011 (06/25/2011) and Sunday, June 26th, 2011 (06/26/2011)
Start Time: 12:00 AM EDT
End Time: 04:00 AM EDT
Services affected: Public / Private Network
Location: GLOBAL
Duration: 28 hours

Event summary:

Datacenter engineers will be performing global maintenance on their backbone. During this maintenance, engineers will be implementing a number of policy changes related to route announcement and accepting routes from providers and peers. This work is required to support continued growth within the network and future global expansion. This is a continuation of the maintenance window from June 18th - June 19th. Due to the scope of this change, the work has been scheduled across the following dates:

Saturday June 25th 12:00 - 04:00 AM EDT
Sunday June 26th 12:00 - 04:00 AM EDT

Customer impact:

While engineers expect no outage during this maintenance, customers may experience higher latency and/or packet loss while routes reconverge across the network.

This is just a reminder that PHP4's will completely be discontinued to be supported by PHP on August 8th, as a result we'll be removing it from our machines August 9th. If you have any concerns whether your scripts will run please contact the vendor of the script, or if you can't get ahold of them you can contact us and we'll take a peak.

PHP4 has numerous security issues and lack of functionality - this is the main reason for discontinuing it in favor of PHP5 - most of your scripts should work and the transition won't be noticeable.

Please see this thread for our previous announcements regarding this.

- Hawk Host Staff

This a topic dedicated to potential issues you could experience due to the migration over to LiteSpeed.

We will be updating all of our servers kernels to the latest version on Tuesday September 21st to to address a recent vulnerability found in some 64bit distributions which affect us. You may recall we run ksplice on all of our servers and they did attend to this issue prior to an official release from the vendor (Red Hat) however due to the nature of the vulnerability we've decided to perform an actual kernel upgrade to ensure this won't affect us in the future.

The maintenance window is for the next 12 hours during which each machine will be updated and experience about 10 minutes of downtime.

Date: 09/21/2010
Start time (EDT): 1:00pm
End time (EDT): 1:00am
Duration: 12 hours
Estimated Down Time: 10 minutes