« 'Accidentally' Upgraded 'box' to LucidGood-bye Gumby, Hello new Gumby »

5-bay RAID and backuppc

05/15/10

  04:20:00 pm, by The Dreamer   , 1373 words  
Categories: Software, Computer, Storage, Ubuntu

5-bay RAID and backuppc

Then Seagate 1.5TB drives came up for under $100 on NewEgg Shellshocker. As always, limit 3 per person....but only needed two more, so I got two. Found that one of the Seagate drives was DOA, and started a full badblock format of the second... where while this was happening, it died and it crashed the port multiplier array causing mdadm to kick two of the 3 disk RAID5 out, that brought 'orac' down hard. Rebooted a couple of times, but couldn't get it all the way up... But, eventually I found the details on how to force it to put things back together and continue the boot after maintenance mode....it accepted two of the disks, and then spent time rebuilding the other disk.

I sent both Seagate drives back to NewEgg for RMA exchange. While I was waiting, I started thinking that maybe I wanted a little bit more redundancy that RAID5....so I looked at options to convert to RAID6.

The gist, was take a drive out of the 3 disk RAID5 (continuing as degraded)...and then use the 3 disks to make a fully degraded RAID6. Make it a new PV, add it to the VG, and do a pvmove.

So, that's what I set out to do....when the two new Seagate drives showed up (after doing a long selftest and a quick format). Unfortunately pvmove didn't work, and it left the VG in a bad state. Again, it left things unbootable...the vg wouldn't activate. Eventually, I found this article -- Recovery of RAID and LVM2 Volumes -- and was back in business.

This time, I made a new VG using the RAID6, and copied the filesystem over. First I was rsync'ng, then cpio....but it was still taking forever...eventually I opted for dd. Things were a bit odd, because now there were two filesystems with the same UUID...and my fstab mounts by UUID. But, eventually I made the old VG and the RAID5 pair go away...and was live on my new, but fully degraded, RAID6. I then added both the drives of the RAID5 in as the 'mirrors' for the RAID6.

I resync'd one of the drives completely first (reading off the 3 data disks), and then started to resync the other drive (using the 4 active disks)...but during this the newly sync'd disk failed....it recovered and continued the long process of resync'ng (from the 3 data disks). Guess I was lucky it didn't go when it was part of the RAID5. But, annoyed that it would be a while longer before I would have a fully operational RAID6.

Not sure which Samsung drive failed, but generally angry about the whole experience...I didn't really feel like doing an exchange, but it was too late to return them for partial refund. So, I just went and ordered 3 more Seagate drives from Amazon.com / Prime.

When that showed up, I put one of them in to replace the failed out Samsung drive...and let it slowly resync. And, then it was done....I debated taking the other Samsung drives out and putting the other new Seagate drives in, but I was leaving for Penguicon the next day...so I decided to cross fingers and hope things survive until I get back.

So, far the RAID6 has been working, though smartd is warning that both the Samsung drives are showing signs of impending failure. One has one unrecoverable sector, and the other has two unrecoverable sectors...also there's a growing count of pending. The array resets now and then as it struggles to recover for bad sectors here and there.

It'll just be a matter of time when I have to replace the other Samsungs....just don't know if I'll do it before they fail or after.

Meanwhile, I'm now considering the purchased of another 5-bay (replacing the 2-bay)...to grow out my storage. Partly because I have other backup / storage desires...and the 2-bay has RAID hardware, and USB support....

But, backuppc has been pretty neat....especially in the recent 'gumby' upgrade. Like it so much, that I set out to have it running at work. Being a Unix shop, we have enterprise back ups of our Unix/Linux systems, so my Unix (Solaris) and my Linux (RHEL4) desktop are included in the backup system. But, I'm the odd duck with a Windows box...I was doing backups to a file, samba share on my Solaris box...but that stopped working at some point, and I never really liked it, and never got around to finding a new solution or fixing it. So, I setup backuppc (in one of my ubuntu VMs on the Solaris Desktop)...and backup my Windows machine. I thought about adding my macbookpro as well, but since it doesn't live at a fixed IP at work....and its on 'open' wireless...I didn't want to have it that exposed for getting backed up. (there's a small hole right now, because at home it has a reserved IP and I know what private IP to allow...) Will just continue using time machine to keep me covered at work....

I did have a separate time machine drive at home, but its kind of tedious switching the preferences between them as I go back and forth....now I can just leave it on the work disk wherever I am, and backuppc can hit it when I'm home....

Average cost per drive $160....hmmm, something didn't work right there....

Pages: · 2

1 comment

Comment from: The Dreamer [Member]  

Oh yeah, it is neat how it says:

* 30 full backups of total size 3464.27GB (prior to pooling and compression),
* 240 incr backups of total size 1072.85GB (prior to pooling and compression).

and

Pool is 1356.68GB comprising 2277643 files and 4369 directories (as of 5/16 01:26)

Should be interesting when I upgrade ‘orac’ from 8.04LTS to 10.04LTS….should at least get pretty graphs ;P in backuppc.

05/16/10 @ 13:51
Now instead of subjecting some poor random forum to a long rambling thought, I will try to consolidate those things into this blog where they can be more easily ignored profess to be collected thoughts from my mind.

Latest Poopli Updaters -- http://lkc.me/poop

bloglovin

There are 20 years 1 month 25 days 21 hours 43 minutes and 45 seconds until the end of time.
And, it has been 4 years 11 months 1 day 16 hours 19 minutes and 11 seconds since The Doctor saved us all from the end of the World!

Search

November 2017
Mon Tue Wed Thu Fri Sat Sun
 << <   > >>
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30      
Google

Linkblog

  XML Feeds

Who's Online?

  • Guest Users: 0
This seal is issued to lawrencechen.net by StopTheHacker Inc.
powered by b2evolution free blog software

hosted by
Green Web Hosting! This site hosted by DreamHost.

monitored by
Monitored by eXternalTest
SiteUptime Web Site Monitoring Service
website uptime