View previous topic :: View next topic |
Author |
Message |
aval Just Arrived
Joined: 28 Mar 2004 Posts: 0
|
Posted: Wed Jan 19, 2005 8:13 pm Post subject: “Why or how do files become corrupt?” |
|
|
Users often ask me this question and I’m not sure how to respond. I did a Google search using the terms above and found one web site that says : “for a variety of reasons”.
Could the following situation contribute or, rather, have nothing to do with it?
One of the servers I manage is rather old (1997 or 1998) and slow:
Single Pentium II 300MHz
512 MB of RAM (should be sufficient…?)
4 4.5 GB hard drives in RAID 5 configuration
RPM? I don’t know. Might only be 7200 (or 5400?).
Now, if multiple users (up to 40 simultaneous) are saving relatively large documents (up to 125MB, although most are around 10-40MB, not counting simple Word docs) all at the same time, if this process takes a long time (exceeding two minutes in some cases), if the hard drive is close to full and quite fragmented…,
could that contribute the file corruption?
I’m trying to get users to delete old documents so I can defrag (not even enough space to run it) but they are dragging their feet.
Personally, I think it’s time for a new file server.
Thank you in advance,
Aval
|
|
Back to top |
|
|
gat0r Just Arrived
Joined: 02 Jan 2005 Posts: 0 Location: Belgium
|
Posted: Wed Jan 19, 2005 8:42 pm Post subject: |
|
|
yes it is time for a new file server. if you dont have enough space to defrag, there will be issues. having 40 people connect to a PII and it not taking a dump is impressive, what OS is it running?
|
|
Back to top |
|
|
RoboGeek SF Mod
Joined: 13 Jun 2003 Posts: 16777166 Location: LeRoy, IL
|
Posted: Wed Jan 19, 2005 8:44 pm Post subject: |
|
|
Lots of things can corrupt files - cross-linking, bad memory, hard drive failing, virus, really just about anything can cause it. Defragging is a good idea, but if your drives are that full you really do need to fix that problem first. - before something bad happens.. like server crashes
|
|
Back to top |
|
|
aval Just Arrived
Joined: 28 Mar 2004 Posts: 0
|
Posted: Wed Jan 19, 2005 10:18 pm Post subject: |
|
|
What OS? Windows 2000 Server.
I say up to 40 maximum. Probably more like 20-25. But there are enough client computers in the school for 40 people to connect to this server at the same time: 2 classrooms with a total of 28 computers, a library with 10 and several computers elsewhere. It's also a DC and DNS server (integrated in AD). Over 300 personal folders are stored on it but, as I stated above, only around 40 users can connect at a given time.
|
|
Back to top |
|
|
geocard Just Arrived
Joined: 19 Jan 2005 Posts: 0
|
Posted: Thu Jan 20, 2005 9:18 am Post subject: try this |
|
|
dear Aval
I would say that it is time for you to get a new server and then do a mirror copy of your old server. to the new one.
This will insure that all your users keep their documents and will insure that you have enough room to play/
|
|
Back to top |
|
|
alpinelegend Just Arrived
Joined: 22 Jun 2004 Posts: 3
|
Posted: Thu Jan 27, 2005 5:46 pm Post subject: |
|
|
I would also like to know just how this server keeps on ticking... What are you feeding it???
Get a new server and FAST
|
|
Back to top |
|
|
liquidz Just Arrived
Joined: 27 Aug 2004 Posts: 0 Location: /dev/null/
|
Posted: Thu Jan 27, 2005 11:08 pm Post subject: |
|
|
One way to have corrupt data, is just an error in the writing or reading of the data that the HD does'nt fix. HD manufacturers will mention something about a 1 in 100000000 probability for a write error in small print somewhere in the product documentation, or on the box.
Older hard drives had a higher probability of writing bad data by "accident", than newer drives.
Last edited by liquidz on Fri Jan 28, 2005 6:25 am; edited 1 time in total |
|
Back to top |
|
|
capi SF Senior Mod
Joined: 21 Sep 2003 Posts: 16777097 Location: Portugal
|
Posted: Fri Jan 28, 2005 12:47 am Post subject: |
|
|
liquidz wrote: |
HD manufacturers will mention something about a 1in 1^10000000000 probability for a write error |
I'm positive you didn't mean to have said "1 in 1" up there... (did you? )
|
|
Back to top |
|
|
liquidz Just Arrived
Joined: 27 Aug 2004 Posts: 0 Location: /dev/null/
|
Posted: Fri Jan 28, 2005 6:27 am Post subject: |
|
|
capi wrote: |
liquidz wrote: |
HD manufacturers will mention something about a 1in 1^10000000000 probability for a write error |
I'm positive you didn't mean to have said "1 in 1" up there... (did you? ) |
That's what happens when you type when you have'nt slept much in the last couple of days. (Worries me what some of my emails and articles might be looking like recently)
Anyway you get the idea, I hope...
|
|
Back to top |
|
|
|