Dart

From Wildsong
Revision as of 15:05, 13 August 2023 by Brian Wilson (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

I sold this system when it ended up powered off >99% of the time. I don't expect I will ever own a computer that bulky ever again.

History

  • 2019-Nov-26 Set up at home and upgraded to Debian Buster (10.2)
  • 2016-09-09 Added intel-microcode package, was getting perf_event_intel: PEBS disabled due to CPU errata, please upgrade microcode
  • 2016-09-08 Upgraded to Debian 8.5
  • 2016-03-27 Lost one of the 3TB drives.
  • 2015-Jul-02 Connected eth1 to Vastra2
  • 2015-Feb-27 -- Bringing online again after months of storage as we moved to California.
  • 2014-May -- Shut down for move.
  • 2014- regrooved with Debian 7
  • 2014- Brought up with CentOS to allow running ESRI ArcGIS Server

RAID commands

My controller is a 3Ware 9260-8i Docs are here: https://www.broadcom.com/support/download-search?pg=Legacy+Products&pf=Legacy+Products&pn=MegaRAID+SAS+9260-8i&pa=&po=&dk=

The controller is 1000:0079, MegaRaidSAS http://hwraid.le-vert.net/wiki/LSIMegaRAIDSAS

The drive is a WDC WD30EFRX "Red".

I have not gotten the RAID command line program running yet on this OS (Debian 8).

At boot time, you can type Ctl-Y when prompted to bring up a command prompt. "$" There is no "help" option -- I found some scant documentation online, here are references

https://supportforums.cisco.com/document/62901/megacli-common-commands-and-procedures

http://www.xtremesystems.org/forums/archive/index.php?t-277864.html

When a drive failed, I had to turn off the ALARM! It's LOUD, you can't miss it. It beeps once per second to tell you a drive is gone.

I have only one controller so I can use -aALL or -a0

Silencing the alarm

AdpSetProp AlarmSilence -a0
AdpSetProp AlarmDsbl -a0

Replacing the drive

There is a walk through on this page: https://supportforums.cisco.com/document/62901/megacli-common-commands-and-procedures

[E:S]  =  Enclosure:Slot

Get information

adpallinfo -a0
encinfo -a0
ldinfo -lall -a0
pdlistt -a0

Mark drive as missing

PDOffline -PhysDrv [E:S] -a0 
PDMarkMissing -PhysDrv [E:S] -a0 
PDPrpRmv -PhysDrv [E:S] -a0

Replace the drive

Rebuild

PdReplaceMissing -PhysDrv [E:S] -ArrayN -rowN -a0
PDRbld -Start -PhysDrv [E:S] -a0

Re-enable the alarm

AdpSetProp AlarmEnbl -a0

Hardware

  • Supermicro server case
  • Supermicro X9SRA/X9SRA-3 v 1.0b 4/18/2012
  • Intel(R) Xeon(R) CPU E5-2650 0 @ 2.00GHz processor (8 hyperthreaded cores, appears as 16 cpu's) 63998.21 BogoMIPS
  • 32GB RAM: 4x KVR16R11D4K4/32 DDR3 1600 Registered ECC (there are 4 empty slots)
  • 240G SSD x 2 for the OS Corsair Neutron GTX (2012)
  • LSI 3Ware 9260-8i with firmware 3.27.00 RAID controller good for 8 drives
  • 7 WDC 3TB 5900 RPM Red disk drives (in slots 1,2,3,4,5)
  • OCZ Revo pci-e card https://www.newegg.com/ocz-revodrive-120gb/p/N82E16820227578
Filesystem      Size  Used Avail Use% Mounted on
udev             16G     0   16G   0% /dev
tmpfs           3.2G  9.0M  3.2G   1% /run
/dev/md1p1      188G  1.8G  176G   2% /
tmpfs            16G     0   16G   0% /dev/shm
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs            16G     0   16G   0% /sys/fs/cgroup
/dev/sda         14T   17M   14T   1% /raid
/dev/md2        112G   17M  111G   1% /var/lib/docker/volumes
tmpfs           3.2G     0  3.2G   0% /run/user/1000

I put everything on SSD, I might be inclined to move /home somewhere.


SSD configuration

The SSD drives are already configured as RAID 1 so I am sticking with that. There is a small (about 200M) unused partition that I probably intended to hold /boot and the rest is 200+G.

OCZ Revo configuration

There is a fast (well it was in 2012, not by 2019 standards) OCZ Revo 55GBx2 card in RAID 0 used for Docker volumes and fast PostgreSQL databases. It's way faster than the spinning drives.

/etc/fstab entry

# OCZ Revo 110GB SSD on PCIe (/dev/md2)                                                                                                                
UUID=3257ae59-d47f-455e-9695-a46c49de68fc /var/lib/docker/volumes  btrfs defaults 0     1

LSI RAID configuration

As noted in /etc/fstab I have 7 drives with 6 in a RAID 5 and 1 hot spare. They are 5900 RPM drives so this is "cold storage".

# 3Ware/LSI raid controller RAID 5, 6+1 HSP                                                                                                  
UUID=34baa852-2bf0-4905-a1d6-f8c6083fc968 /raid           btrfs   defaults        0       1

Network

  • eno1/br0 192.168.123.?
  • eno2 unused

Software

  • Debian 10
  • postfix not exim4
apt install docker-compose net-tools dnsutils btrfs-progs emacs postfix

Postfix

Email is supported via Wildsong at the moment. That is, mail originates @wildsong.biz

  • smarthost using gmail, because I use gmail for the Wildsong and Geo-CEG domains.
  • sasl auth is needed to talk gmail

Mediawiki + MySQL

DEPRECATED

I used to install these here. My wiki moved to the cloud long ago. If I did install here it would be Dockerized.

PostgreSQL

I run PostgreSQL and friends in Docker now. The pain of PostGIS installation on the host outweighs any minor performance gain. I do have to make sure the data ends up on the correct logical drive on the host!

DEPRECATED

Anyway if I wanted to install on the host it would require adding a repo to get a more recent version. In days of yore the commands would have been

echo "deb http://apt.postgresql.org/pub/repos/apt/ wheezy-pgdg main" > /etc/apt/sources.list.d/pgdg.list
wget --quiet -O - https://www.postgresql.org/media/keys/ACCC4CF8.asc | apt-key add -
apt-get update 
apt-get install postgresql-9.4

2015-Feb-27

I packed up Dart last May (2014) shortly after re-installing the operating system and switching to Debian. Now I am trying to pick up the pieces. There is no Web server right now! Zow! How can that be?

I do remember I gave up on CentOS when I found out how impractical it was to try to run my own copy of ESRI ArcGIS Server after I lost (free) access to a developer license.