[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [leafnode-list] Merging 2 instances of Leafnode
Mark schrieb am 2004-02-10:
> >What leafnode versions are you running?
Too old. Please update to 1.9.49 or 1.9.50.rc4.
> >If it's leafnode 1, do the spool directories have the same name length?
> >/var/spool/news and /usr/spool/good are evidently same name length,
> >/leafnode and /usr/local/var/spool/news are not. This is important for
> >the hash function.
> The spool directories are the same length.
In that case, identify (reading manual pages) a tool that can preserve
hard links even for a large number of files and across directory
Then stop all cronjobs that might affect leafnode.
Now make a backup of the new spool, just in case something goes wrong.
Copy copy your old spool into the new spool, omitting the .overview
files. "rsync -aH" will work, recent tar implementations should be fine
as should be pax or star. Avoid cpio unless yours handles "ustar"
After that, make some spot checks with "ls -l" in the copied directories
whether the link count is >= 2 for the copied all-numerical article and
If and only if all spot checks have a link count of 2 and higher for
article files, everything is fine, run texpire to correct the groupinfo,
then reactivate your cron jobs.
Note that when you're using leafnode 1.9.49 or older, the globally first
access to a news group will be very slow.
WARNING: if the spot checks reveal files that have a link count of 1,
DO NOT RUN TEXPIRE UNLESS YOU HAVE 1.9.50.rc4 INSTALLED - it would delete
all your articles. Update to texpire 1.9.50.rc4 first, then run texpire.
If after that run with 1.9.50.rc4 you still have files with a hard link
count of 1, please report the file names, then move the spool out of the
way and restore your backup.
Encrypt your mail: my GnuPG key ID is 0x052E7D95
leafnode-list mailing list