FreeBSD Upgrades - Sunday 01/30/05 8:00PM EST

[allera]

Notice sent to all customers:


Dear Customer,

We are scheduling a maintenance window this coming Sunday, 1/30/2005 at around 8:00-9:00 PM Eastern Time (GMT -05:00). During this maintenance window, we will upgrade FreeBSD from 4.10 to 4.11.

All unix services will be affected except for the control panel, which runs on a Linux server. We anticipate 10-15 minutes for the upgrade for each server.

Should you have any questions and/or concerns regarding this scheduled maintenance, please do not hesitate to contact us at support@fluidhosting.com.

Please plan for this maintenance window accordingly. Thank you for your cooperation.

Sincerely,
Administrators
Fluid Hosting, LLC


The SHARED hosting Control Panel server runs Linux, so it will be unaffected.

The RESELLER hosting Control Panel server runs FreeBSD, so it will be affected.

[KrisB]

Hey guys... is that a 10-15 minute outage, or 10-15 second outage expected?

If the former, could you please move this to the 'wee hours'? We're talking about 5-6pm PST, which is probably a bad time to be down...

Thanks,

-kb

[FH-Dave]

KrisB,

That would be 10-15 minutes, but we expect it to be much shorter.

Unfortunately, there is no 'wee' hours that fit everybody since we have global customers all around the world.

[FH-Dave]

Another clarification email has been sent to all shared hosting/reseller customers.


Dear Valued Customers,

This is just a clarification on the scheduled maintenance notification sent earlier.

Just a reminder for all of you, the scheduled maintenace will be performed on Sunday 01/30/2005 at 8:00 PM - 09:00 PM EST (GMT -05:00). The following FreeBSD servers will be briefly affected for the kernel upgrade during this maintenance window:
- All mail servers
- All MySQL and PgSQL servers
- All Unix web servers
- Forums, IRC, and support.fluidhosting.com
- ns2.fluidhosting.com and ns3.fluidhosting.com, ns1.host80.com, ns2.host80.com.

The following servers will not be affected:
- Control Panel server
- MSSQL server
- All Windows web servers
- ns1.fluidhosting.com
- Windows Web servers

Should you have any questions/concerns, please do not hesitate to contact us at support[at]fluidhosting.com.

Sincerely,
Administrators
Fluid Hosting, LLC

[KrisB]

Hi Dave,

I recognize that, however as a North American hosting company, I would expect that the majority of your customers are on this continent.

I would also be surprised if there isn't a large part of Asia that doesn't have many customers at all for you -- it just seems exceptionally odd to have this scheduled at what is definitely primetime for anyone in North America (between 5pm and 9pm, depending on what part of the continent).

Cheers,

-kb

[inTELLiGrunt]

Hi KrisB, that mailie definately said 15 minutes, not 15 seconds. If you're that confused over time how can I be sure your not equally confused about space and cyberspace :D Which continent are you on? That timing is fine by me :)

[KrisB]

intelligrunt: I am aware of what it said. In the past, upgrades that are scheduled for 15 minutes were about 15 seconds in duration, which is why the question was posed.

I'm not sure of the difference between 4.10 and 4.11, and I really don't feel like checking out the ChangeLog. It could be a recompile kernel & reboot situation, which makes it a 15 second outage... but if it means installing many new binaries then it could take 15+ minutes.

Given the use of time ranges and approximate downtimes, the request for a more precise estimate was warranted, I believe.

And, it shouldn't take much more than a 3rd or 4th grade education to understand from my previous post that I'm in North America. Especially considering the fact that the "Location" section of my profile is filled in and each post shows it as "Prince George, B.C., Canada" :rolleyes:

Fine with that timing, you may be... it's not something that I would see as worthwhile at this point... there's no major holes that have been fixed (just checked the ChangeLog) -- and that is very poor timing for pretty much all of the Americas.

[inTELLiGrunt]

KrisB: It's nice to know that you were aware of what you said and that you have finally explained why the question was posed.

I'm more realistic, I know that boxes don't go down and come up in 15 seconds. To me the mail was clear, some time during a one hour window I can expect outages on UNIX servers. Even when the mail was clarified, the message was the same; the timing of which suits me fine.

I know what continent you're on, I was wondering whether you knew, it was a more localised version of the "what planet are you on?"; in a joking sense. :D

[allera]

The time was chosen specifically because our techs would be in the datacenter working at the same time. If anything goes wrong with the upgrades, there would be hands immediately available.

Realistically, the downtime should only be a few minutes (under 5) at most, due mainly to a reboot. However, we'd like to have the option to run a few minutes longer if needed. This is why in the past it has been stated in minutes when it really only took seconds.

Those of you who monitor the servers closely will likely notice the downtime, but anyone who doesn't have sensitive monitors probably won't.

Play nice. :)

[allera]

The server upgrades are taking a bit longer to do than anticipated. The window is being extended to 10PM. There really shouldn't be more than a few minutes of downtime.

[allera]

The upgrades are being postponed because of unforeseen compiling problems. Mail1.fluidhosting.com is also having problems that are being taken care of.