faq
search
add article
cool-stuff
how-to
main
tips
parent
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jennifer Zhao on Tuesday August 26, @06:55PM
|
Hi,Jason and everyone
I have been using this drivecopy.sh for a couple of months to do full backup once a week. It worked very fine till yesterday.
Last night, As usual,I started ./drivecopy.sh to backup my FreeBSD4.3 server (Disk1) on another HDD(Disk2). Disk2 has been used successfully as backup for more than 10 times with same procedure of ./drivecopy.sh.
So it seemed the start up going smoothly, then I left the server room.
This morning, when I came to check the server and backup. (which was supposed to have finished). And I surprised to find all those horrible error messages on the console screen:
***start to quote error message
ad0s2e: had error reading fsbn 97103692 of 3473456-3473467(ad0s2 bn 97103692; cn 6044 fn 108 sn 28) status=59 error=40
cannot read: BLK 3473456
UNEXPECTED SOFT UPDATE INCONSISTENCY.
**end of error message quote
So I had to force to suspend the backup process by pressing Ctrl+Z.
Then I tried to reboot Disk1. I had problem during the booting process.
**start to quote error message
....
chflags: not found
chown: not found
cd: can't cd to /var/run
cd: can't cd to /var/spool/lock
/etc/rc: cann't create /var/run/dmesg.boot directory nonexistent
cd: can't cd to /var/run
/etc/rc: cannot ceate
/var/run/clean_var:dirctory nonexitent.
Enter full path of shell or return for /bin/sh
.....
**end of error message quote
Then I tried my luck to boot Disk2.Unfortunately, it showed the same errors as Disk2 during booting.
Can anybody tells me what I can do to recover those two HDD, Disk1 and Disk2?
Your any advice will be highly appreciated.
Jennifer
P.S. The attached is my drivecopy.sh,which had worked fine for the previous backup processes except yesterday. So I do not think anything wrong with the drivecopy.sh itself. There must be something wrong with my yesterday's backup process, though I can not find out what it is. |
|
|
The Fine Print: The following comments
are owned by whoever posted them.
( Add a Reply )
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jason on Wednesday August 27, @10:07AM
|
***start to quote error message
ad0s2e: had error reading fsbn 97103692 of 3473456-3473467(ad0s2 bn 97103692; cn 6044 fn 108 sn 28) status=59 error=40
cannot read: BLK 3473456
UNEXPECTED SOFT UPDATE INCONSISTENCY.
**end of error message quote
Hi Jennifer,
It looks like your original data drive has a problem with the partition which contains your www data. The error states the FreeBSD could not read a block of data off the original drive. I have seen this error caused by a troubled IDE cable as well as a bad removable drive bay. If you don't suspect either of these issues, then most likey your IDE drive is starting to fail. The unfortunate part is that this error occurred during the time of your backup- which means you have NO backup now.
I would try to drivecopy the known bad partition by itself to see if you can squeak one in so to speak. The problem could be thermal, so you may be able to let the drive cool and then attempt a backup. Chances are good that you have a successful backup of all the other partitions because ad0s12e was the last partition in your drivecopy shell script.
You may also want to boot to single user mode and run an fsck on the www partition first.
Let me know how it goes and I'll try to assist you in any way I can.
Cheers,
Jason
Note: If you have remoevable drive bays in your system, I would take your drives out of the bays and attempt bypassing those bays for the backup. I would also replace your IDE cable(s). It probably won't resolve the problem but it's inexpensive and it certainly won't hurt.
|
[
Add a Reply to this ]
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jason on Wednesday August 27, @10:14AM
|
Jennifer,
I didn't read your post carefully. You definitely will want to boot in single user mode and run fsck -p
|
[
Reply to this ]
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jennifer Zhao on Thursday August 28, @11:58PM
|
Hi, Jason
I run fsck -p, it still show that some bad block can not be read.
So I suspect it is hard drive problem. Since the hard drives are still under warranty time. So I took the IDE HDD to the supplier. Their technican used a special software, something like Maxtor Blast, (which is provided by the HDD manufacturer of Seagate or Maxtor to test the HDD problem). And their technican said the test show no problem on the HDD.
So, I am just wondering when the fsck show the bad block error, does that mean hardware problem of HDD or just corrupted data, or either?
Best regards,
jennifer
|
[
Reply to this ]
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jennifer Zhao on Wednesday August 27, @11:51PM
|
Hi, Jason
First thanks very much for your kind advice.
It seems it is not the problem of IDE cable or thermal. For I waited it cool down and use a working IDE cable, and the same error comes.
And I also enter into single user mode and issued
#fsck -p
The result is the known bad partiton show the same error message:
ad0s2g had error reading fsbn 14498.... The following file system had UNEXPECTED INCONSISTENCY...
So, it seems the fsck -p doesn't work here.
Anyway, I tried the way you said"drivecopy the known bad partition by itself". But it seems I can not do it. For when I issue the command #mount -a, I got error message:
Warning, R/W mount of denied. File system is not clean-run fsck...
Anyway, my guess is that there are some bad partitions in my two drives (IDE HDDS) starting to fail. And the data on them are not retreiveab le any more. A little luck of me is I have a third drive at home, with everything same as the bad drives, except the data on it was 1 month old, for I used it for a full backup one month ago.
Based on those experience, I have a couple of questions here
1.Is it worth to use those two bad drives again and reinstall a simple system on it and drivecopy.sh everything from the third drive to it? Or just discard those bad drives, in case cause my third working drive fail together?
2.How many time of drivecopy.sh can be used on a new HDD before this HDD worn out and can not be safely used any more?
3. I guess a weekly back up by using drivecopy.sh is a little bit too hard on HDD.
But I would express here that drivecopy.sh is a brillant method to move whole FreeBSD system from one drive to another. No doubt at all. The only concern here is whether there is any limit times you can utilize drivecopy.sh on a HDD.
4. Any recommandation on what medium to use to back up a small business server, with not big traffic database (mysql) and a website hosting.
Best regards
Jennifer
|
[
Reply to this ]
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jason on Saturday August 30, @09:44AM
|
Hi Jennifer,
First, I don't think a weekly drive copy is asking to much of your hard disks. How much data are we talking about? If it's only a couple of gigabytes then you should expect your drives to hold up just fine.
I've personallly seen your "fsbn error(s)" on two drives of my own. On one the drive was bad, no question about it. On the other, it turned out to be the removable drive bays I was using. The drive stopped having problem as soon as I moved it out of the bay.
I have a good contact with a computer store and they see drives that will fail in a system but will pass the manufacturers required tests. The problem is that computer stores often cannot return a drive under warranty if it passes the manufactueres silly little tests. Case and point; I had a problem with a Dell hard disk that could easily pass Dell's 9090 test but would fail constatntly when in the system. After bantering endlessly with Dell tech support I finally just bought a new drive, copied the data over and the problems went away. Dell never replaced their malfunctioning drive even though it was under warranty!
Re: Question 1-4
1. I would do whatever you can to save your data.
2. Many times unless you have very large, very full drives. I've been doing a monthly backup on two pairs of drives since about 1999 without a problem. I recommend single user mode for drivecopy backups on a system whos data is changing regularly (i.e. an SQL server).
3. drivecopy.sh is a reliable way of moving an installation and doing periodic backups (ie not daily).
4. If you need to do regaular daily backups I would use a SCSI DAT drive (I prefer Seagate). I have a Samba server at a site where I use drivecopy.sh whenever a major change is made to the system, but use the DAT drive to backup their data on a daily basis.
Hope this helps.
Cheers,
Jason
|
[
Reply to this ]
|
Re: HOWTO: Move FreeBSD to a new hard disk
by Jennifer Zhao on Sunday August 31, @03:40PM
|
Hi, Jason
Thanks a lot for your detailed reply, which gave me confidence what to do next.
I agree with you that manufactueres' silly little tests on their HDD doesn't gurantee their HDD at all!
In fact, Yesterday, I used the following commands and seems it recovered the corrupted partition, though not the data, which I think it has lost anyway.
--I bootup the troubled HDD, and enter into single user mode
--issue command
#mount ad0 /mnt
#newfs /dev/ad0s2e
#shutdown -h now
--Restart the computer, this time the booting was alright, and I used fsck -p, it seems all the partitions are clean now.
Maybe this is the work around way to fix the troubled HDD.
Just hope it will NOT happen corrupted data during backup in the future.
Thanks to recommand Seagate backup tape.
However, since I do not know much about Seagate backup tape, do you mind I ask you a couple of questions about it?
1. Which one you recommand: Internal or External one?
2. which Technology I should choose for FreeBSD backup, DDS-4, DDS-3 or DAT72 ?
3. Which interface I should choose: SCSI-2 or ULTRA 2 SCSI LVD
Your advice are highly appreciated
Jennifer
|
[
Reply to this ]
|
|
The Fine Print: The following
comments are owned by whoever posted them.
( Article Reply )
|
|
Related Links 
