| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
hi, |
| 5 |
|
| 6 |
|Well, using a Raid 1 in that instance probably isn't the best |
| 7 |
solution. Raid |
| 8 |
|1 arrays are really best used in a critical server to keep the system |
| 9 |
online |
| 10 |
|until a tech can schedule maintenece. |
| 11 |
|
| 12 |
what do you mean by say that 'raid 1 arrays are really best ...' ? |
| 13 |
'best' ? why ? |
| 14 |
|
| 15 |
| Soo many things can happen to a raid 1 |
| 16 |
|that it is not a good solution for data backup. If a file is deleted, |
| 17 |
|corrupted, etc it is just mirrored to both drives. There is no |
| 18 |
rotation, or |
| 19 |
|history. |
| 20 |
|
| 21 |
i haven't ever considered this to be a data backup |
| 22 |
the really important things like mailbox or sources i'm working on, |
| 23 |
have regular tape backups, etc .. |
| 24 |
the reason why i have raid 1 on my workstation is, that if a drive |
| 25 |
crashes, i still want to be able |
| 26 |
to work, while waiting for the drive replacement to arrive |
| 27 |
additional 160gb drive cost me less then 80 eur and if it saves me 4 |
| 28 |
hours of time which i wouldn't |
| 29 |
be able to work without having raid 1, it was a good investment |
| 30 |
|
| 31 |
the fact that everything is 'just' mirrored to both drives is exactly |
| 32 |
the reason why i'm using it |
| 33 |
no matter how good your backup strategy is, if we are talking about a |
| 34 |
workstation you are using |
| 35 |
10 hours a day, you will always loose something |
| 36 |
i won't, unless both drives crash at the same time |
| 37 |
|
| 38 |
|Unfortunatly, its a lesson not commonly learned until it is too late |
| 39 |
:( Raid |
| 40 |
|is not a replacement for backups. It does help is a lot of |
| 41 |
instances, it is |
| 42 |
|not a bad idea, but it is flawed. |
| 43 |
|
| 44 |
|
| 45 |
as i've said, i don't think it is a replacement for backups |
| 46 |
i just take it as a quite good protection against loosing couple of |
| 47 |
hours just because of a disk failure; |
| 48 |
and if i am talking about couple of working hours, it is a replacement |
| 49 |
which pays for itself, given the current |
| 50 |
prices for ide drives |
| 51 |
|
| 52 |
|and create snapshots of the system configs and user files. On my |
| 53 |
main server |
| 54 |
|for my business, I dump the database and rsync every 30 minutes, and |
| 55 |
rotate |
| 56 |
|
| 57 |
|
| 58 |
on our e-commerce server, i replicate production database every 10 |
| 59 |
minutes to another server and have tape backup |
| 60 |
every night |
| 61 |
but this is completely different topic |
| 62 |
|
| 63 |
it is all about the cost of you backup strategy (now you can also |
| 64 |
include hardware backups, etc ...) vs. outage costs |
| 65 |
i was talking about a workstation, where you can invest up to 100 eurs |
| 66 |
to almost eliminate a risk of a drive failure |
| 67 |
|
| 68 |
|backup. The question of the day is how active is your data, and how |
| 69 |
far back |
| 70 |
|can you recover from. If losing a hour of work can kill you, then you |
| 71 |
|probably back up more often. Most people are fairly safe daily. On my |
| 72 |
|database, 30 minutes could be upseting, hence why it is so anally |
| 73 |
backed up. |
| 74 |
|
| 75 |
in my case, loosing of 30 minutes on the db server could upset |
| 76 |
customers who haven't downloaded results yet, but would |
| 77 |
definitely upset us because of inability of charging our customers for |
| 78 |
these 30 minutes |
| 79 |
that's why we do db sync every 10 minutes and think of getting raid 1 |
| 80 |
there as well; but in this case it is much pricey option |
| 81 |
because of scsi prices |
| 82 |
|
| 83 |
loosing of an hour of work can't usually kill you, but loosing 3-4 |
| 84 |
hours of work may be more expansive then additional 160gb drive |
| 85 |
for raid1 on a workstation |
| 86 |
|
| 87 |
|risk involved, then that is your choice. Just remember when you |
| 88 |
recommend |
| 89 |
|solutions to someone that may not be familure with the pro's and |
| 90 |
con's that |
| 91 |
|you let them know. Nothing is worst than a user that feels they are |
| 92 |
backed |
| 93 |
|up only to realize after the fact that important file is lost forever. |
| 94 |
|
| 95 |
|
| 96 |
if you recall the beginning of this thread, it was not a question |
| 97 |
about recommended backup strategy, it was a question |
| 98 |
about installing lvm over raid on scsi drives |
| 99 |
|
| 100 |
if i ask the question, i expect to get an answer to the question |
| 101 |
asked, not a recommendation why should i do things differently |
| 102 |
|
| 103 |
if i ask a wrong question, it is my fault and i should know better |
| 104 |
|
| 105 |
but of course, this is just my point of view |
| 106 |
|
| 107 |
best regards, |
| 108 |
martin |
| 109 |
|
| 110 |
-----BEGIN PGP SIGNATURE----- |
| 111 |
Version: GnuPG v1.2.4 (GNU/Linux) |
| 112 |
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org |
| 113 |
|
| 114 |
iD8DBQFAa3GTFCvm4cvhqkwRAoj0AJ4pCGP7+q1oRqUbgs5Y1338Ld180ACcDItA |
| 115 |
mmQvyo8aChgyvU7p4F+Fp3E= |
| 116 |
=19G/ |
| 117 |
-----END PGP SIGNATURE----- |
Archives