Platform: RH 6.0 on i586 (K6-2 350) Problem: INN's inndf tool, when invoked normally (no switches) reports "0" bytes free in whatever directory it's pointed at, despite inndf -i (reports free inodes) reports plenty of spare inodes, and a regular df reports (on my system) over 70MB free. This in turn causes innwatch to throttle the server immediately. It's reproducable even after the recent security-fix release, and I have noted the behaviour on the binary i386 version with released RH 6.0 and the 2.2-9 rpm release built from the src.rpm example. [news@cartman news]$ inndf /var/spool/news/articles 0 [news@cartman news]$ inndf -i /var/spool/news/articles 101000 [news@cartman news]$ df Filesystem 1k-blocks Used Available Use% Mounted on /dev/hda5 565887 467054 69603 87% / /dev/hda1 1822884 1478132 344752 81% /mnt/win98 TIA, Michael Fleming.
Fixed (by using f_bsize rather than f_frsize as returned from statvfs) in inn-2.2-12.
*** Bug 3466 has been marked as a duplicate of this bug. *** inndf always reports 0.
*** Bug 3797 has been marked as a duplicate of this bug. *** The binary '/usr/bin/inndf' should return the free space in Kb when called with a path. But it returns ever zero ('0'). So the script 'sendbatch' doesn't work because this zero return value. BTW: the update from RedHat 5.2 to 6.0 doesn't work correctly for inn if it couldn't remove the '/usr/lib/news' directory.
*** Bug 3841 has been marked as a duplicate of this bug. *** This program is supposed to replace the df | awk... but it is not working. The problem is that this program is used by other scripts (including sendbatch...). I've replaced it by : #!/bin/sh df -k $1 $2 | tail -1 | awk '{ print $4 }'
*** Bug 4023 has been marked as a duplicate of this bug. *** It looks like you compiled with -DHAVE_STATVFS and not -DHAVE_STATFS. As it is now it always returns a "0" for the inndf. I have gotten the source and recompiled with -DHAVE_STATFS and it works.