[HECnet] A2RTR downtime

Robert Armstrong bob at jfcl.com
Mon Oct 18 08:19:40 PDT 2021


>Second is VENTI; I had wanted to test something on it and it doesn't appear to be online.

  Sorry, I’m somewhat lazy about booting it up when the host system is rebooted.  I don’t use VENTI that much.

>Had you had any interest in trying it?

  It sounds like I’m going to have to.  It’s a shame this STOPCODE was a halt and not just a debug one; at least the latter would have allowed the system to continue running.

  This also means I have to re-discover how to do a MONGEN and build.  There was a time, 30 years ago (actually, 40 years ago) when I could have done it from memory but now I’ll have to read the manual again.

>Were you aware of any winning backup solutions?

  Partitions are your friend – keep all your files separate from the system files.  For the non-root partitions I have shell scripts that back them up (using tar and bzip2) over NFS to other servers.  I also have an external hard drive that I can plug into the system and do a backup to that drive whenever I feel the need.  The hard drive lives in a fire safe when it’s not being used.  That’s not 100% safe against physical destruction, but it’s the best I’m willing to do.

  The root partition I usually save with dd (and bzip2).  This partition is usually pretty small (a few tens of GB is plenty for a running Ubuntu system, even with a bunch of stuff installed) and doesn’t take that long to copy with dd.  If the system disk dies or is destroyed I can buy another disk, create an identical root partition, dd the image back, set up GRUB2, and it should boot right up.  The I can use tar to restore my other partitions at my convenience.  

  I’m sure there are other ways to backup your Linux system, and there are probably tools to automate some of these steps, but it’s really not hard with a few scripts and standard utilities.  

  RAID1 is protection against the failure of a single drive (assuming you have two drives in your array) and that’s about it.  It won’t save you from anything else.  This system has two physical SATA drives and, if I set up the RAID correctly, they should be exact copies of each other.  That includes the GPT and GRUB, so that if the primary drive fails then you can boot directly from the second drive w/o changing anything. 

Bob

 

 

  I can restore Tops-20 to bare metal, but I really don't remember how to do this for Unix (although I did know it for Ultrix at one point).  So I starting looking.  Déjà Dup looks like it won't quite do what I need, but since it uses duplicity, I started looking at that.

Remember, even a RAID is no substitute for backup.  This was probably more true in the days where a hardware RAID controller error introduced a single point of failure; it may still be true for a software RAID.


  _____  


On 10/17/21 12:03 PM, Robert Armstrong wrote:

  A2RTR will be down tomorrow (Monday) for a couple of hours sometime between 9AM PDT and 3PM.  It’s not my doing – PG&E is replacing a transformer and the power will be off.  I have battery backup for some of my equipment, but not A2RTR.  Sorry.

 

  It’s also going to have to go down sometime again, maybe next weekend, so I can replace a disk drive that’s failing.  I bought two identical drives and plan to set up a RAID 1 array this time.  If anybody has set up a software RAID on Ubuntu, using the md driver and mdadm, for the system drive and you have advice, let me know.  Espcially if you managed to make both of the drives bootable.


Bob

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20211018/d384f8a2/attachment-0001.htm>


More information about the Hecnet-list mailing list