FAQ Profile
Search Members
Groups

PM's

Register Login/Out
Concerning Email Problems on March 3rd
Last Thread | Next Thread  >

Post new topic Reply to topic
Your-Site Members Forum > Your-Site System Status

Author Thread
Guest





Concerning Email Problems on March 3rd  Reply with quote  

Let me start by appologizing for not getting anything posted earlier. I do most of my work in the evenings, starting around now, and I'm afraid I wasn't able to get hear earlier today although I would have prefered to have this posted first thing. I know everyone has been waiting to hear from us and I'll do what I can to address the major questions.

First, concerning credits and refunds: We will give our standard 1 month ($7) credit for downtime because of what happened. I haven't had a chance to confirm if this is being done somehow automatically to every account. It couldn't hurt to send an email to support@your-site.com requesting the credit, and I encourage everyone to do so. I will try to find out if we are going to do this automatically for every account, but I still encourage you all to proceed.

If you decide to cancel your account refunds will work as follows:

* Cancelling within the first 30 days will give you a full refund.

* Cancelling after that will give you a pro-rated refund based on your remaining time.

Second, concerning the RAID failure: As many people know when a drive fails on a RAID system you should be able to swap it out, swap in a new one, and even recover the damaged data. As someone already stated, RAID systems are designed not to fail. I believe we have swaped bad disks from RAID units in the past with no problems.

This time when we tried to replace the drive and resynch the unit we repeatedly encountered errors and the resynching could not proceed. We exhausted every method we could try to resynch the drive before we were forced to start from scratch. That is what caused the loss of the email.

If you want further information on this I could ask our Sys Admin if he can provide anything else.

Third, concerning backups and storage: Mail is currently stored on a 500 GB RAID volume. For our size this should be quite sufficient. For anyone who may be unclear, this is not the same as having everything stored on one hard disk. If your C: drive in your desktop died you'd be in a lot of trouble. In this case 1 drive in a unit of drives failed. We should have been able to replace that and recover all the information due to the way RAID drive works, but that goes back to my last point.

As we have mentioned we are planning a massive server upgrade and the new mail system should be running by the end of the month. We will have redundant RAID units for all systems and are looking into ways to work this for the email system.

We do not do realtime backups of email messages. With the new system we may be doing nightly backups of any messages, but we are still examining the feasability. Doing realtime backups of email messages would have at least triplled the load on the email server. Anyone who has complained about the current server speed should be able to imagine what kind of effect that would have had. The RAID unit was supposed to act as a backup mechanism, being able to recover any data should a disk go bad. Several posters here have such faith in RAID units that we have pretty much been accused of lying about what happend they find it so unlikely that such a unit would fail. Unfortunately that is what happened and so our backup system failed. We did have backups of everyone's email settings and accounts.

I should note that we have always maintained webmail was not originally intended as a standard email client on our system, but as a temporary client when you are away from your desktop. Under this idea your email would generally be saved on your computer and not our system. If this were the case then little email would have been lost yesterday. Before I sound like I'm trying to pass the buck please let me say I realize how popular web mail is with many users and that many of you did have important email saved on the server. We were pretty late in installing a web mail system compared to other hosts - indeed I don't think we have totally settled on what system we want to permanently use at this point - so I suppose our system is still geared more towards email clients that download and then delete mail off the server. I'll reiterate that we are examining our options for changing this with the new system and better backing up email messages. (and of course the RAID system should have recovered the data/shouldn't have failed, etc. Sounding like a broken record yet? Yes, I think so.)

I should mention web sites are also backed up.

Fourth, concerning our responsibility: Ok, I'm sure you're all sick now of hearing "The RAID system shouldn't have failed!" It did, and that's why we're hear today. However, I will be the first to say we are the ones responsible for maintaining and providing functioning hardware. While the failure may not have resulted from any action or inaction on our part we are still the ones responsible. (I'm not sure if there were any signs of this coming, although I know some people have pointed to things like the pop server not properly disconnecting after checking email and receiving multiple copies of single messages. I'm not sure if those were related to either the bad disk or the RAID's (impending) failure. Instinctively I think no, but I can follow this up with Eddie or the Sys Admin if you want.) We do take full responsiblity for this.

Let me add no one has ever heard a Your-Site employee say, "You get what you pay for," or "What can you expect for only $5 a month?" Those phrases are oft-repeated by some around here, but never a Your-Site employee. Certainly that is not the philosophy either Eddie or myself approach this job with. If another Your-Site employee, and remember Eddie and I are the only employees currently posting here, has said such a thing to anyone here please let me know and I'll try to have the matter looked into.

We do greatly regret what happened yesterday and appologize. If there are any points I have missed or any further question you have please let me know. I would like to try to keep further posts about this topic to this thread to make it easier to gather information in case I need to ask Eddie or the Sys Admin any more questions.
Post 03-05-2004 02:12 AM
 

Post new topic Reply to topic
Forum Jump:
Jump to:  

All times are GMT - 4 Hours.
The time now is 09-21-2018 07:50 PM
  Display posts from previous:      


Powered by phpBB: © 2001 phpBB Group
Icicle template by Vereor.