At least they're honest

B

bread_man

Guppy
The recent downtime had a huge affect on my website and organization. Although I understand how such things can happen, I was dissapointed that us customers had to suffer as much and that there wasn't a better system in place to minimize downtime and get customer websites up sooner.

However, since I've been a member, the staff at JH has always taken full responsibility for their actions and thrown themselves at the mercy of their customers via email and in an open forum. Nothing is more frustrating than when you have a host lie to you about problems and you have no way of knowing what's going on, and no recourse. These guys don't seem to operate like that, and its appreciated. Honesty counts.
 
I am satisified that after the IIS metabase failure, whatever we did to recover, we put every single option on the table and tried each and every one of them. I don't think we could have done any better

However yes, this problem shouldn't have happened... We are going to be very careful henceforth with hsphere patches are working to have our new backup system in full swing even faster (allowing faster disaster recovery).

This issue has been the most stressful I've ever faced, more stressful than any other issue I've encountered here at JodoHost.
 
I agree...they've been honest.
You'd think they would've learned from the past problems to be ready for install issues and have an immediate solution.

I wrote Yash today regarding the down-time and still wonder why they don't have, at the very least, a spare server to host the sites.

Aside from that idea...they could also use clustering.

It still boggles my mind and those I referred to Jodo, which is why we're looking elsewhere at the moment.
 
I don't think there is anything called an immediate solution.
In any immediate solution, you need to still update customer files and that is whats taking the longest at the moment

We'll be announce our new strategy to handle such issues after we are done with Win5. Trust me, we won't let such an issue happen again. I have your email, I'll be answering it a little later
 
Yash, in April you posted "Sure... I do not take anything as an attack against us. We had these 2 failures on Win5 we accept it and take responsibility for it.

We use a system called RAID1 on our Windows servers and raid5 on our linux servers. RAID1 and RAID5 are disk arrays that can sustain the loss of a disk without the loss of any data. For example, our win servers have 73GBx2 SCSI Hot-swap hard disks. x2 mean that there are two hard disks on the system. These 2 hard disks are exact mirrors of each other. This mirroring is done by the help of a RAID1 controller from adaptec. If one hard disks fails, there would be no data loss and the system would continue to run.

Apart from this system, we take full system backups every 3 to 4 days and move them to a remote server for safe storage.

We plan to introduce a new system that would help us to increase uptime. This system is based on a secondary server concept where we have a secondary server to take over if the primary server (for example win5) fails without any downtime or data-loss. A program would run on the primary server that would monitor the primary server for any file changes and immediately update with secondary server with the latest files. Therefore, the secondary server would theoratically have an exact mirror of all data on the primary server at all times, and would be ready to take over from the primary server at any time, automatically

We are hoping we can deploy this system in the middle of June or earlier"

In August you posted about new backup procedures at http://support.jodohost.com/showthread.php?t=2644 which talks about hot-swappable disks including Win5.

You actually mentioned "Through our new backup strategy, we keep a complete mirror of the server's hard disk on a separate SCSI disk which is kept off-site. If a server goes down with a OS or system crash, recovering the server would simply involve swapping the current hard disk with its mirror." after commenting "The main issue with these forms of backup is recovery. After a system crash or OS corruption, we have to prepare a fresh system. Recreating the resources of over 1000 sites on a fresh standby server can take a good hour or two and then restoring the data can take another 6 to 8 hours at least (although the entire process is automated)."

More recently another post explained backups to be on a remote datacentre. So you have been accutely aware for quite some time of the need for backups and speedy data-recovery as close as possible to immediate.

The question is why the August plan was not implemented as this would have minimised downtime as explained in that post and all the agony?
 
eko, EVERY ONE of those backup strategies are outlined

I already mentioned about the hot-swap backup as in the August plan. Look in the Win Q&A backup. That worked on only for 30 minutes before it started crashing as well

We had backups of all customer files... so this is nothing like the last crash. But yes, the actual issue shoudn't have happened and we take responsibility
 
Yash, first of all I quoted you in these posts, but please talk to Stephen and ask him what he needed to do to get my site back up. I'll eagerly await your reply as what you maybe believe is happening is not based on what has happened.

We also seem to differ on the interpretation of "immediate".
 
eko, I am not denying whatever happened is a totally unacceptable to customers and we are at fault. I am just trying to answer the specific question on these issues

Yes, I am aware of your issue. We did have a backup of your account, it was unfortunately overwritten by a script and we had to recover from an older one (the older one can be 1 to 3 weeks old, only the current set is within 24 hours )
 
Yash we possibly interprete your statement "We had backups of all customer files... so this is nothing like the last crash." differently.
 
You must log in or register to reply here.
Back
Top