You were hosting on a server with only one HDD?. Are you mad?. What type of Hard Drive was it?.
-
Howdy! Welcome to our community of more than 130.000 members devoted to web hosting. This is a great place to get special offers from web hosts and post your own requests or ads. To start posting sign up here. Cheers! /Peo, FreeWebSpace.net
Time to rebuild my host spread
- Thread starter TaoPhoenix
- Start date
You were hosting on a server with only one HDD?. Are you mad?. What type of Hard Drive was it?.
Rounds
Actually, you are doing the right thing by holding off until things stabilize. The "rounds" are loose, mostly when some kind of mini epoch seems to complete itself. Fasterhost got stuck with the worst possible glitch right out of the gate, but I think I've dug him out of "out of fire only into soup". There are ZERO benefits to aggressively entering early - it's not a "day counting contest". Hosts only earn creds racking up months of "No Bad News Today". If you get stuff squared away and you really think you have it locked securely, you may still be able to enter Round 3. There aren't that many tiers - there isn't much difference between a host entering in February and one from April.
Actually, you are doing the right thing by holding off until things stabilize. The "rounds" are loose, mostly when some kind of mini epoch seems to complete itself. Fasterhost got stuck with the worst possible glitch right out of the gate, but I think I've dug him out of "out of fire only into soup". There are ZERO benefits to aggressively entering early - it's not a "day counting contest". Hosts only earn creds racking up months of "No Bad News Today". If you get stuff squared away and you really think you have it locked securely, you may still be able to enter Round 3. There aren't that many tiers - there isn't much difference between a host entering in February and one from April.
Seraphim
Active Member
500GB Western Digital. Failed 20 days out of the box. This is why you always make backups. Though usually such a drive will either fail in the first month of operation, or will survive several years without incident. Unfortunately for me this one was the former, and I was at least prepared for it.
Heck most of the dedicated servers available out there only have a single drive in them unless you pay extra. Either way, they're installing two hard drives in RAID1 as a replacement, so it won't matter anymore unless my luck is that bad that I have both of them fail on the same day.
I still have the backup from when I moved to that machine in the first place, the only major change was an update to my own site scripts and those are all edited off the server and uploaded anyway.
Either way, I had been bug tracking the new server for some time now and holding off on entering until all of the issues were resolved. It seems that the true source of the issues has been found in any case, faulty hard drive certainly explains the quota errors when I tried to run test accounts to prove all of the features.
I'll let you know when I am good to go again, I doubt it will be more than a week unless I get more problems.
..Green drive by chance?.
Seraphim
Active Member
Blue they tell me.
I've just been notified that my server is good to go again, so I'll be setting things up as soon as possible.
Heh I posted so many messages this weekend the user status thingie flipped over into "FWS Addict" range!
That's lovely, a host that doesn't even know their own hardware configuration.
If your server was using a lot of disk I/O that drive was getting hammered and probably would fail. I use blue myself, but at least I use RAID10. Blue is not a black drive, and they are not really designed for server use.
You should go with a raptor drive for a single disk server, or a RAID configured system. RAID1 in my books doesn't cut it unless your just using it purely for redundancy. In my honest opinion you should be looking at RAID5 or RAID10.
I hope that you take measures to continue backing up your data at any rate as when it comes to redundancy with RAID, the only thing that it will protect you from is a disk failure. If data gets corrupted it still gets corrupted on both drives.
Seraphim
Active Member
Drive shouldn't have been getting hammered upon in a new build that for the most part was sitting very close to idle. And yes I am going to continue routine backups, that's a given. the RAID is just for better performance and to reduce the probability of this happening again.
RAID1 wont increase write performance.
Backups
Hi Colorhost.
(Giddy) Backups are a good thing. I think you should do lots and lots of backups! Like once an hour! (/Giddy)
Is there a downside you know of in site performance when you do backups? It's great that you're being responsive.
Hi Colorhost.
(Giddy) Backups are a good thing. I think you should do lots and lots of backups! Like once an hour! (/Giddy)
Is there a downside you know of in site performance when you do backups? It's great that you're being responsive.
colorhost
Member
Somtimes when i backup Performance does go alittle down, but not by much.
It may act up if the client that i backup has alot of things in his account, it just depends on whos account is backing up at the moment.
PS:I wont do the Hourly Backups! Hourly will kill the server constantly! I do Weekly though.
Thanks
ColorHost
Last edited:
My general principle for the study is that minor slowdowns aren't damaging. Only major slowdowns are a cause for concern, because that's the warning sign of either oversellers or tech difficulties.
Seraphim
Active Member
Well I am sure having a good time over here.
One of the drives in the RAID1 is showing warning signs, I may have to tear down again for yet another drive replacement. Though this time it won't be as long of an outage, since a physical replacement and array rebuild is faster than a clean setup.
We'll see how it goes, as long as it stays to bad sectors that will eventually be flagged and not used anymore I should be done testing and go ahead. Hopefully no other problems will surface.
I can understand all new drives having some bad sectors, but it would be just my luck that I hit 47 bad sectors on a drive brand new out of the box in the first three days of operation.
Either way, all of my services are back up I am just tuning the configuration and testing everything.
One of the drives in the RAID1 is showing warning signs, I may have to tear down again for yet another drive replacement. Though this time it won't be as long of an outage, since a physical replacement and array rebuild is faster than a clean setup.
We'll see how it goes, as long as it stays to bad sectors that will eventually be flagged and not used anymore I should be done testing and go ahead. Hopefully no other problems will surface.
I can understand all new drives having some bad sectors, but it would be just my luck that I hit 47 bad sectors on a drive brand new out of the box in the first three days of operation.
Either way, all of my services are back up I am just tuning the configuration and testing everything.
Last edited:
Status Update Round 3
Happy Easter.
Here is the current status of Round 3.
http://taophoenix.paradoxservers.net/ReVision.html
We have a situation where some hosts are direct resellers of Paradox, so there's a bit of a domino chain effect lurking, but they seem to have solved their local problems with unrelated service glitches.
I have two more slots available, and then it we can all go back to "just using" the hosts for ... the pages we host with them. In other words, the Round 3 excitement seems to be over, so according to my rough observations, we'll have to wait for about 4 months for the newcomers to pass the first duration metric.
Happy Easter.
Here is the current status of Round 3.
http://taophoenix.paradoxservers.net/ReVision.html
We have a situation where some hosts are direct resellers of Paradox, so there's a bit of a domino chain effect lurking, but they seem to have solved their local problems with unrelated service glitches.
I have two more slots available, and then it we can all go back to "just using" the hosts for ... the pages we host with them. In other words, the Round 3 excitement seems to be over, so according to my rough observations, we'll have to wait for about 4 months for the newcomers to pass the first duration metric.
It looks like Seraphim and Lost Signal will be closing out Round 3. It will take me a few days to process out the signups, but that's the formality side.
I've been overhauling my design again too. That link I posted is getting near a usable form, with the iPhone being a key testing metric. (Though I don't know how to spin it sideways, and it's not a priority at all. I have wasted years procrastinating, so I have to finally buckle down and start researching the articles.)
The site shell structure will be done soon enough, perhaps a couple of weeks, and in the medium-soon future, it will be rolled out to synch the copies among the hosts. But even though the concept is to write the articles in layers, expect it to be months before anything substantive actually appears. Unlike Microsoft, I despise Vaporware announcements.
Happy Easter,
--Tao
I've been overhauling my design again too. That link I posted is getting near a usable form, with the iPhone being a key testing metric. (Though I don't know how to spin it sideways, and it's not a priority at all. I have wasted years procrastinating, so I have to finally buckle down and start researching the articles.)
The site shell structure will be done soon enough, perhaps a couple of weeks, and in the medium-soon future, it will be rolled out to synch the copies among the hosts. But even though the concept is to write the articles in layers, expect it to be months before anything substantive actually appears. Unlike Microsoft, I despise Vaporware announcements.
Happy Easter,
--Tao
Paradox Glitch?
Hey Darksoul,
Any news on this? I can't connect to your server.
----
Address Not Found
www.paradoxservers.net could not be found. Please check the name and try again.
Hey Darksoul,
Any news on this? I can't connect to your server.
----
Address Not Found
www.paradoxservers.net could not be found. Please check the name and try again.
