Originally posted by: RedSquirrel
Originally posted by: Goosemaster
Originally posted by: RedSquirrel
At least you *HAVE* backups. Where I work nothing really gets backed up properly. It's brutal. I've basially been doing exchange backups manually every now and then with windows backup. They take about 5 hours, and it's stored on the same SAN.
But yeah, I hope we never need to use the backups...
This makes me realize how the new backup solution needs to be put in place STAT. Too bad there's so much crappy politics in IT, anything takes forever to actually get off the ground. Servers are so much easier at home. The day I wanted a home backup solution, I had the equipment ordered, and the day it arrived, I had it setup. I don't get why big companies like to lag projects so much. There's planning, and then there's screwing the dog.
dear lord... you have to drive the message home. have numbers and situations ready to present that show how absolutely fucked they would be if things were down, much less unrecoverable.
I, like everyone else, face budget limitations, but made the most of what I had (Audit, then GFS as soon as possible, D2D2T soon thereafter, encryption, optimization etc.) and still have a ways to go.
You backup that stuff using ntbackup to external drive if you have to. Hell I had to do that (I call those the "bad times). The most basic requirement is to have SOMETHING and work up from SOMETHING to SOMETHING DECENT to SOMETHING USEFUL and so on
<---wants bare metal but not at that stage yet
edit: this assumes that we aren't talking about individual stores that are larger than a terabyte in which case you're fvked
Right now it's decent, the entire 4 stores put together are a bit under 200GB so they actually fit on a single DLT2 tape. Problem is I don't know enough about the tivoli backup system on how to even do a manual backup. I need to find some really good doc, read it, know it, and redo the system or something.
But yeah, it's bad when my home backup solution is better then the office one.
so we are in the same league then
I would definitely make sure you have something reliable and consistent that works. ntbackup (2k.2k3) isn't sunshine but you can backup most anything somewhat reliably with it. TSM is pointless to waist time on if you haven't put in a solid 0-day solution in place with or without it. Get an old box, stuff a bunch of disks into it with freenas or openfile and just have a nice ntbackup incremental/differential and full backup that overwrites every week if nothing else. Odds are you can find the spare parts in the office to build it and it will be worth it's weight in gold.
Then, and only then, fiddle around with TSM.*
*= so jealous
Originally posted by: RedSquirrel
Originally posted by: hiromizu
Reading this thread makes me realize that what we've got in place is never overkill.
Nope, it's better to have too many backups, then not enough.
in some situations it can drive you insane but what supersedes too many vs not enough is the business need. you have to analyze what is worse: immediate downtime prevention/recovery, archival restoration ability, complete DR(fvk!), and if it isn't clear you need to plan accordingly.
if it's not cut and dry, having a backup implementation suited for a non cut and dry set of situations is what matters. otherwise you are just wasting space
Originally posted by: adlep
So the Exchange store is down?
Will not mount?
Maybe I can help...
Have some expertise with recovering stuff.
1. Make sure that you have enough space on the partition where all of your stores reside.
2. ESUTIL is your friend.
3. Offline defrag is a must at the end.
4. Tape sucks. What you want to migrate to is product called Backup Exec System Recovery 8.5 which is basically a Norton Ghost for Servers. That + External USB 2.0 hard drive will get you a nice recovery point that you can restore back to your partition in no time.
2. this particular issue made the store unrecoverable. trust me we tried.
4. I have no server that will be able to keep with our tape (lto-4). I have to "tone it down" to 80MB/s with restores coming in a only a few notches lower. There is also the durability factor*, the perm-offsite factor*, the the cpu usage**, among other things.
Needless to say, despite a lack of snapshots, my system is pretty decent. We are looking at very short down time for full IS failure and I am working on improving that everyday.
*=disks usually allow you to pile on backups on the same media. From an efficiency standpoint you get massive capacity and respectable speed at a basement price with off the shelf components. the density is quite a liability when you realize that a single drop could take out everything though. the fact that constant usage of these discs requires an on site presence is a no go as well. I have had a transitional period where we did D2D2D but that was only temporary. unless we are talking rugged disks (those that can handle a 1m drop on concrete), I wouldn't risk it unless you have to. there is also the fact that tape is EXTREMELY reliable if you treat it correctly. I have no qualms with staking my reputation on tape.
**=usb, especially local usb, is very processor intensive and when you add software encryption and compression it gets worse. the fact that tape drives are usually parallel scsi or sas means that the controller is doing a lot of the work for you, minimizing the cpu usage.