nt38 11/28/05

[badchauncey]

When you get past the letter of my user name, does that mean that I'll see the code back in the folders of the domains? Right now it is just the generic "control panel" stuff from hsphere when I hit the domain. Just curious

[Brangwyn]

Permissions also need to be reset once restores have completed, basically I wouldn't count on anything working correctly until Tech's say the entire process has completed. Though of us that have been through a few disk failures now and then I think have just learned to wait it out, when they say it's done, thats when things will be working again.

[Silverbug]

Any updates on this? what letter / % are we up to now?

[dickkyboy]

I have just been told by support that it will be 6 to 8 hours more. ARGH!

At least for the 99.9% uptime we are going to have 85 years of uptime without anymore downtime. Nice!

[dvanburen]

Both the data restore and the account recreation are at roughly %50.

[supportsolution]

Hi,

Following this thread I have noticed that the server is having some php problems then it was fine. Then it is doing a check-disk. This is run from boot up of the OS so the drive is not "locked". Why would the server be rebooted outside of a maintenance window?? was the first comments by tech misleading ?

If the server started a check-disk it would suggest the OS is intact to launch a check-disk. A check-disk would only start if it told to, or the server power is terminated. Servers don't just decide too without reason.

If your server got to a point where it would not let you logon, did you try and boot in "safe" mode ?

If the OS was intact as it started to do a check-disk and it had slowed down so much it never allowed you to log in, why did the disk array break - that suggest the array was intact, but the os was at fault.

If the array is not rebuildable, who can say there is no data lost?

The array should not take long to rebuild a raid 1 array. You should be able to make the mirror drive the "master drive" and a "mirror drive" should be a new drive. - this should have made downtime minimum. (initialise the new drive, rebuild on the fly when the OS is up)

Redundancy as in replicating server, not as in hardware. If we are talking hardware then I assume all servers have redundancy memory, CPU and PSU.
A raided server is pointless if the whole system is not redundant. (RAID 03/5 have disk performance benefits as well as redundancy benefits)
If you have a raid back plane problem, what happens to the content of that server? while you get a new back plane?

Don't get me wrong, I am not having a go. All I have had from you guys is fantastic service, but I think you have found a skills gap that needs fixing asap. Also a redundant server that can be fired up so downtime is kept to a minimum.

Why did I post? - I used to look after cluster servers with drive arrays of over 30 hard drives for the stock market in the London, Casino's in Australia, manufacturing companies and the G8 summit in 2000. These servers where allowed 1 hour of downtime a year. I know my high availability servers and in a way web servers are high availability servers. Please guys treat them that way (because we do) and get proper redundancy in place. Disaster recovery plans are created for these situations, it really seems that you guys misplaced your document.

Not loosing trust in you lot - keep up the good stuff.

Dave.

[dvanburen]

I'm sorry, I don't have time to address all of your points, but I would like to point out that BOTH drives were replaced, and data restored from backups. This is not a rebuild of a degraded array.

[dickkyboy]

Here Here... Nice post supportsolution.

This is a disasater for those with critical websites on this machine - we have a mall of computers showing NTXX instead of their website. Terrible Terrible situation and it is going to cost me alot of money and clients.

The thread doesn't make sense as has been demonstrated in this previous posting.

If we are talking a software fault - can we be assured that a restoration is going to fix the problem or are we going to be looking at another days downtime?

Oh I have a headache.

[supportsolution]

Hi,

What restore software are you using ? It is slow as! ... Not NT backup... Please

By anychange, you were not using Maxtor SATA drives where you?

D

[Silverbug]

If the server started a check-disk it would suggest the OS is intact to launch a check-disk. A check-disk would only start if it told to, or the server power is terminated. Servers don't just decide too without reason.

I belive earlier on they mentioned that after they had rebooted the server it started a check disk

By anychange, you were not using Maxtor SATA drives where you?

Im pretty sure they use western digital 15k scsi drives

This is a disasater for those with critical websites on this machine

I agree that this is a disaster for anyone with sites on the server. But if your sites are that critical, maybe you should look at hosting them somewhere that accomidates those sorts of website. I for one have a number site I host elsewhere, on servers where uptime is essential (and obviously the hosting cost reflects that). To be honest we get a damn good deal when it comes to the price we pay for the service we receive here.

[dickkyboy]

silverbug, most of the time i'd agree about your comments of support. But when it takes two support calls, 45 minutes and 8 emails to be sent a link to this thread, it makes one re-think that.

All because they don't attach a support id to your support login. Or prompt you for it.

[Silverbug]

Hmm yeah i see your point too. However where would you rather the tech's be, fixing the servers or answering support questions where people are asking how the fixing is going? it would be nice if they tried to make a point of updating the forum on the hour, just to reasure everyone. Even if its just to say no changed from the last update, but we are still working on things. However saying that i would rather the problem just be fixed, and if that means we not getting told till its done then so be it.

[kapuwa]

keen to know on which letter we are now? i have a "s" customer almost ready to send me a bomb

[admin]

We only use seagate hardware these drivers are SATA 250 GB drives http://www.seagate.com/products/ente...aces.html#SATA this was also our first SATA system as a test system with raid1 well, less then 6 months we see how well that went over. SCSI rocks..

This post is not miss leading in any way. A post is to update customers as techs have info to post. This post was first made for the system rebooting and it came back online a bit latter is when it died. We just kept the start and end all together along with the new issue of the drives tkaing a leap.

supportsolution are you even a customer? If so please PM me other wise c-yea.. You have 2 post both here..

But to answer some of your questions NO we do not use NTbackup. I have to ask why does every one thing moving 100GB of data with over 100 of 1000's of files in 8 hours or less slow? Do me a favor take a good web site and copy it 1000 times then try and move that site accross a network. These are not single file sites sites have from 100 pages/images or other stuff to over 10,000+. It's not something that can happen fast a drive just can't read that much less write that fast.

Also supportsolution your servers a server for "tock market in the London, Casino's in Australia, manufacturing companies and the G8 summit" is a lot different and set up a LOT different then a shared hosting server is. A lot of people think a server is a server just because they call it a server. Sure it might be windows, sure it might have IIS, but I bet that IIS has 1 to 5 sites. I could restore 1 to 5 sites by hand in 10 min. but if the content is 1000's and 1000's of files it still takes time to move.

We have taken steps to use raid and every time it has done nothing but really cause us more problems then a fax or a save for anything at ALL. I have yet in 6 plus years running vortech really seen raid1 save anything, raid5 I have seen a few time. But then again no one here pays enough for us to raid 5 100+ H-Sphere system. Again this comes down to what people are willing to pay for and well why are 90% of our customers here? Because we offer an above avg. solution for below market price along with top notch support and real people you can get on the phone and speak ENGLISH unlike some others out there.

We are doing everything we can to restore this system and there is no way to make H-Sphere or the restore of the content go any faster. Just because we are telling you the letter does not mean your site will come back on line when we say the letter here is why.

H-SPhere is running the recreat for IIS/Users and other info to setup the sites. So we can't copy you user data in to c:\hshome because H-Sphere would over write it with the default data because it does not go from A - Z it is random. Our restore goes from A - Z so the letter we are telling you is content only because content in most cases take longer then H-Sphere. In this case it's going to be a very close race to what gets done first content or IIS.


supportsolution "Also a redundant server that can be fired up so downtime is kept to a minimum." well you are again talking about a LOT of money and well there is no way to do it on $35 per customer buget sorry just not going to cut it. I like to eat. Also this is not really the issue NT38 and it's OS was back online on a new system with in 30 min of getting the new drives in and kicking the old ones out because they were indeed died. We wanted to give the old drives some time and work before giving up on them and tossing them out for a new system as we know this would have been the best for all of us.

supportsolution, I am not picking on you I just don't know you and your not shown as a Vortech Inc. customer and it's hard to tell from your 2 posts and you just signed up.

[dickkyboy]

Is this serious, or just a joke from a non-customer?