On Fri, Apr 26, 2024 at 12:44:34PM +0200, Peter J. Philipp wrote: > Hi! > > I've had some problems with dump(8) lately. A 800 GB SSD partition on a > raspberry pi 4b (via USB) that is 50% filled had trouble with dump. I don't > know why this could be, but it used to work. > > Here is my backup script that I used to run in my "nodump" chflagged > /home/pjp/Backup directory. Notice the old behaviour, which is hashed out. > For some odd reason the not so large dump's have seemed to make it. > > -----> > #!/bin/sh > > umask 027 > dump -0ua -h 0 -f - / | gzip -c > vega-root-backup.dump.gz > dump -0ua -h 0 -f - /var | gzip -c > vega-var-backup.dump.gz > #dump -0ua -h 0 -f - /home | gzip -c > vega-home-backup.dump.gz > > find /home -type f -print > filelist.txt > find /home -type f -size +8000000000 -print > excludelist.txt > fgrep -v -f excludelist.txt filelist.txt | cpio -oz -H pax > > vega-home.cpio.pax.gz > > echo These files were excluded from packing with cpio format pax: > cat excludelist.txt > > exit 0 > <---- > > I actually haven't run this script yet since I did all of these commands > manually on the command line but they are 100% compatible. > > Best Regards, > -pjp
There was an issue with sigsuspend unlocking and that got reverted but snapshots have not caught up to that yet. Since dump(8) is a heavy user of sigsuspend it tends to deadlock on big partitions. -- :wq Claudio
