[SOLVED] Services dying on snapraid sync execution!

jmgibson1981 · 04-03-2021, 06:29 PM

*EDIT* Solved here https://www.linuxquestions.org/quest...8/#post6237249

Using one of my parity disks as a temp holding just in case. Something is off though. I did an rsync from my /snapraid/pool (mergerfs directory) > this mountpoint.

Code:

250gb:2tbspinner:500gb:750gb:thick2tbusb  4.4T  2.9T  1.4T  69% /snapraid/pool

Code:

/dev/sde1                                 3.6T  2.3T  1.2T  65% /snapraid/parity/4tb35

How can I have 2.9t total on the pool, but rsync only moved 2.3? I know my content files on each aren't that big and overwrote each other. 600gigs isn't a small error.

Code:

rsync -auv --progress /snapraid/pool/* /snapraid/parity/4tb35/
sending incremental file list

sent 4,670,106 bytes  received 3,178 bytes  718,966.77 bytes/sec
total size is 2,427,928,040,391  speedup is 519,533.60

*EDIT* Just checked, no hidden directories / files either. Could I possibly have managed to duplicate some stuff at one point? I have had a few drives go south so in copying data off the drive... idk. Only answer?

computersavvy · 04-03-2021, 07:33 PM

1. If there is matching data already at the destination that rsync is copying it will not be transferred. Rsync is different than cp in that cp copies everything and overwrites what is already there, rsync only copies what is needed to make source & destination match (thus the "sync" part.)

In fact with the command you posted just above, rsync only sent ~4.5 MB, not 2.4 TB

jmgibson1981 · 04-03-2021, 10:49 PM

I understand that. I ran that command that I posted to show there was nothing missing or synced and still the discrepancy.

computersavvy · 04-04-2021, 12:19 PM

I cannot answer further since I have no clue about snapraid and using different sized devices in the raid.

Maybe "df" can give you a little more clue with its numbers of total, used, and available space.

jmgibson1981 · 04-04-2021, 05:18 PM

Ok. Going to adjust the title of thread to why this whole mess started and explain.

Original title for posterity "Migrating snapraid > raid 1 + lvm. Inconsistency with data totals?"

First off, 99.9% sure all my data is there and I had duplicates. The reason I was looking to switch to raid 1 + lvm was ever since I removed that drive my services would crash on snapraid sync, namely ssh and docker. This is obviously a huge issue. I never found anything in the logs indicating why this was happening during a sync. More frustrating is I had recently moved the server to a remote place with no screen or keyboard. So I've carried it across my yard once or twice... Big yard. As such I cannot definitely with logs prove this is the reason but here is where I'm at.

Snapraid doesn't like duplicates apparently. Near as I can tell I was basically giving my machine a heart attack when I told it to sync. In theory duplicates should never happen if done right. I did not. Instead of restoring lost disk from parity I moved what I could off it into the pool and managed to make a major duplication. As of 3 hours ago when I started the sync my ssh and docker services were at 18 hours of uptime. I had set a crontab for both to restart if they died which runs every 2 minutes as I could be gone for awhile. I just got home, 3 hours later and my ssh and docker are now at 21 hours up. Sync is still going and I had immediate ssh access when I connected.

Ultimately the problem is solved and I now have zero reason to switch to lvm & raid 1. To be honest I probably never would have found this out had I not decided to make that switch once these problems started happening the other day. I didn't bother to post about it because I needed it up now, not whenever I got help so I just did what I could to get it up asap.