• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Project: Rackmount Overkill

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
The seller of the Brocade switch didn't bother to reset the administration passwords before selling it, which is kind of a hindrance. He claims he did not have them, so my options were to try and reset the password, or return the unit. I've been waiting quite a while for one priced this well, so I figured it was worth a shot trying to reset it. Problem is, Brocade support is probably closed and probably wants money, so I figure I'd take a shot at resetting it. The reset utility while booting the switch requires you to contact Brocade, so no-go there.

The switch just runs a stripped out Linux, so working with it is easy. I found a few hints online that showed how to feed it boot options, and located one that dumps you right to a root shell. The problem is knowing what address on the hard drive to boot. Either way, good as new:

brocade_200e_configured_1.png

You can see the green light on port 1. It is connected to Ruby.


brocade_200e_configured_2.JPG


Still trying to find rails for the switch. The ones that came with it were certainly not a full set. I also need to do some wire management.
 
Last edited:
Worked on the R710 tonight, trying to get the new RAID card working in it and to get the array migrated over. I would really appreciate that when people sell stuff on eBay, specifically computer hardware, that they test to make sure it is in working order before selling it and certainly before sending it. I'm half referring to the switch above and half to this RAID card. The card "somehow" arrived with a completely discharged battery and was so discharged that the RAID controller thought there was no battery when I first powered it on. After a few minutes charging, it finally realized it did have one, but it kept throwing "Battery or memory" errors. I'm unsure why it couldn't differentiate between the two, but I figured it was related to the battery. The fix was to pull the BBU off the card, boot and tell it there was no BBU, shut down, then plug it back in with a mostly charged battery. It seems to be ok now, but I'll test it more tomorrow.

I was going to copy over the data with Clonezilla, but it somehow doesn't have the drivers for that server's network card. It was late enough that I'd had enough of working on it, booted it normally to continue folding in the Chimp Challenge, and went to bed. Sleep time.
 
The 8708EM2 is finally running in the R710. I had issues getting it running, noted above. Today, I put the network card in so Clonezilla wouldn't have a problem and made the image of the existing install. I then shut the server down, swapped the cables, and fired it back up. The RAID card did not complain about anything being wrong and reported the BBU as working fine. I created the RAID 10 array remembering that it has an absolutely stupid setup procedure for this level, but got it working after a few tries. After it says that "continuing will blow your disks up", it tells me the BBU isn't connected, working properly, or going through a relearn cycle.

WHAT. No it isn't. It is connected, fully charged, and working. The only thing in the logs that I see is that it reports it wants a recharge cycle to get more buddy-buddy with the charge levels. Whatever. I'll let you think about what you've done while I re-image the disks. Jerk.

So I fire up Clonezilla to restore the disks, it goes through the restore painfully slow, gets to installing GRUB and just stops. I give it a few minutes, thinking it is detecting disks and the RAID controller is throwing a tantrum. Nothing. Odd, that usually is pretty quick. I jump to another tty, run top and see it is maxing out a core. Huh? Ok, well that is pretty weird, I'll give it a few more minutes. Five minutes go by, still nothing. Thinking something has to be wrong, I do a quick search to find out that the version of Grub that Clonezilla has won't work with CentOS' GRUB because it is on EXT4. There is a message that flashes by that you are supposed to magically notice (I carry my high speed camera on me at all times, I don't know about you) that states "THIS DOESN'T WORK WHY ARE YOU DOING THIS", which then continues to run and doesn't work. I don't get why it doesn't, you know, stop. But hey, not like I'm a programmer or anything.

The solution is to get the Ubuntu version of Clonezilla. Fine, just give me the files, I'll do it. The developer specifically notes that you have to have the same architecture as the installed version, so I need the 64bit copy, ok, cool, let's do this bro. Go to the download page, grab the ISO, burn it, boot it. The OS gets to where the language selection should be, but all I see is a random splattering of pink mist pixels on the screen. Then it turns green. Dude, I must have got a Dell, this thing is tripping hard. Hit the magic buttons on the keyboard to reboot and it does it, so the disk just hates my video controller, good information. Sit through another restart and select the "seriously my video card is a flaming pile of Voodoo2" option and it comes up fine. Restore the image, everything goes through eerily well, system restarts, OS comes up.

I look at the clock on the wall. I then realize there is no clock on the wall and look at the time on the computer and see it is four hours later. That was a giant waste of time.

Walked up to my desktop, fired up a terminal, installed the MegaRAID manager on the R710 to monitor the battery cycle, and I think I'm just going to sit here staring at my monitor until the screensaver comes on.

r710_8708em2_1.png

r710_8708em2_2.png
 
Last edited:
Aww, the card alerted me. It wants some attention.

r710_8708em2_4.png

What do you want, little guy?

r710_8708em2_3.png

You have low self esteem? I can't really help with that.
 
Testing out ZFS with my 1 TB drives, at the moment.

Code:
[root@ruby ~]# zpool create -f zfstest raidz sdc sdh sdd raidz sdf sdi sde
[root@ruby ~]# zpool status
  pool: zfstest
 state: ONLINE
  scan: none requested
config:

        NAME        STATE     READ WRITE CKSUM
        zfstest     ONLINE       0     0     0
          raidz1-0  ONLINE       0     0     0
            sdc     ONLINE       0     0     0
            sdh     ONLINE       0     0     0
            sdd     ONLINE       0     0     0
          raidz1-1  ONLINE       0     0     0
            sdf     ONLINE       0     0     0
            sdi     ONLINE       0     0     0
            sde     ONLINE       0     0     0

errors: No known data errors
[root@ruby ~]# zpool list
NAME      SIZE  ALLOC   FREE    CAP  DEDUP  HEALTH  ALTROOT
zfstest  5.44T   202K  5.44T     0%  1.00x  ONLINE  -
 
That's what I guess. That's the reason why I'm building such an extensive home network - to play around with it and get familiar. I would much rather trial-and-error on things at home rather than do them at work.

Same with getting certifications - you can read books until you are blue in the face, but for higher level certs you really do need hands on experience.
 
can I get a legit answer please?
I've answered it multiple times in the thread. The file server is a file server. The others are virtual machine hosts, which run a variety of VMs: databases, camera monitoring, web servers, full Windows networks, etc. When I find something that is interesting, I test it out in a VM and keep it if proves useful.
 
Preliminary tests with ZFS on the 1 TB disks is still happening. I'm testing out different features and settings, such as compression and de-duplication. So far, I'm really liking how it works and performs. I let a copy of files run last night with ZFS iostat running every 30 seconds, which I've attached below. I copied data until it threw an out of space error. Considering it didn't even get to the sequential media files, the ~300 MB/sec it held for minutes at a time is pretty impressive. I'm doing a test with compression now and it is doing ~110 MB/sec while loading about half the processors on the server.

I still need to do resilver tests and make sure this M1015 works with the SAS expanders.
 

Attachments

  • zfsfilecopy_raw.txt
    40.5 KB · Views: 75
Probably, but 300 MB/sec on my main storage array is going to be faster than the network accessing it. That figure is also while writing, which might be limited by the current main array. Reading might be quicker. I haven't tested that yet.
 
If you want to do a good read write test you can use dd.

write test:
dd if=/dev/zero of=tmp.dat bs=1024k count=2x the amount of ram in system

read test:
dd if=/dev/tmp.dat of=/dev/zero bs=1024k count= same as above


You can even do random write / read test using /dev/random instead of zero.


I used the above to test my zpools when I was using zfs.
 
Back