Disaster Recovery and Backup Advice
- Thread starter Jakeee
- Start date
The only DR & BC project I've been involved in resulted in us virtualising all of the servers onto a Hyper-V platform. We used Backup Exec to do a full daily backup of the entire infrastructure to tape, which was then sent off site. We had two alternate sites provided by a dedicated BC provider with pre-allocated servers waiting for us to need them. The idea being that we get the tapes shipped from the off-site storage facility directly to the BC site, then restore the entire environment from tape onto the Hyper-V servers at the other end.
We did a trial run of the restore process, it took us a day to get the entire environment up and running, then left the data restore running overnight. By the following morning, we had an exact replica of what was running back in the office.
Obviously that's a pretty high level description of what suited the specific organisation, and probably wouldn't be a good fit for many others. One of the biggest limitations was that the tapes we were restoring from were from the previous Friday, which means potentially up to a week of data might be lost.
As sidethink said, you've left this pretty open so it's hard to answer in any more detail - I assume that's what you had in mind, though?
We did a trial run of the restore process, it took us a day to get the entire environment up and running, then left the data restore running overnight. By the following morning, we had an exact replica of what was running back in the office.
Obviously that's a pretty high level description of what suited the specific organisation, and probably wouldn't be a good fit for many others. One of the biggest limitations was that the tapes we were restoring from were from the previous Friday, which means potentially up to a week of data might be lost.
As sidethink said, you've left this pretty open so it's hard to answer in any more detail - I assume that's what you had in mind, though?
We tend to run on the cheapest possible solution.
We take a full backup + diff of each server to 3 offsite NAS (located 100 meters down the road).
Backups keps for two weeks. Important Database backups Full every night and logs every 15 minutes to a standby server and NAS.
Does us ok, need to test the backup strategy at some point....
Old place was D2D2T offiste backup and fireproof safe, Gransfather-father-son.
We take a full backup + diff of each server to 3 offsite NAS (located 100 meters down the road).
Backups keps for two weeks. Important Database backups Full every night and logs every 15 minutes to a standby server and NAS.
Does us ok, need to test the backup strategy at some point....
Old place was D2D2T offiste backup and fireproof safe, Gransfather-father-son.
Always a good plan.
A few years ago, I was looking through some servers for a new customer we picked up. Nice rack full of ProLiants with somebody changing the backup tapes every day. One problem : was there no backup job configured (let alone scheduled) in Arcserve...
I tell this story quite often, a local council got torn a new one during an audit for not having any IT BC planning in place. After procuring much budget they had a massive project to get all sorts of measures in place. The grandest being a diesel generator to kick in when the power fails.
They tested this generator religiously every month and were very confident in their systems they'd invested in. That was until an actual power failure and they realised there was hardly any diesel left in the generator from all the testing!
Enterprise level, maximum safety.
Currently have the following: All servers are running raid 5, we have 2 daily backups: 1 for exchange and SQL databases and the other for work documents and archive files, these are stored in a safe, off site and go back 2 months before recycling. (Every month, I take out of the 2 tapes and store them - so we always have a copy of on month - this goes back 6 years).
And before long, out exchange server will be hosts via a private cloud and connected via a leased line which will be mirrored and managed by another company.
Anything I should change or improve?