On Tue, May 24, 2005 at 10:11:28AM +0200, Samuele Giovanni Tonon wrote: > Ross Boylan wrote: > >Package: apcupsd > >Version: 3.10.17-2 > >Severity: normal > > > >The distributed killpower script does the first two mounts without a > >remount, and the others with a remount. This looks as if it might be > >an oversight. > > > no, this is done on purpose. > This shell script if placed in /etc/apcupsd > will be called by /etc/apcupsd/apccontrol before > apcupsd kills the power in the UPS. > > apcupsd kills the power in the ups after that the kernel > is umounting filesystems . please see > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=301916&archive=yes Then why do the other mounts in killpower use remount? If the filesystems are unmounted, remount should be unnecessary. > > > > >If that's not an issue, it would probably be appropriate to downgrade > >this to "minor" severity. > > > >All my other issues concern the documentation of killpower. > > > >The file doesn't really indicate what environment the script will be > >run in, making it hard to assess issues such as that raised above. > >Have all the partitions been unmounted when it runs? > you have to know what apcupsd is doing when it calls killpower > maybe the documentation is not very clear, but it is true that > you should take a look to see how and when killpower file is called, > by who, and which conditions happens when it's called. > > This could be explained on the killpower file, however it's > self explaining if u look at apccontrol file, which you should > look if u want to configure correctly apcupsd. I didin't know that. > > > > > >The Debian NEWS is confusing. It says > > Also there's a new file in /etc/apcupsd which is called > > killpower, > yeah and that's the file we are talking.. or not ? Yes, it is. > > >if you have /usr/ and /var on different > > partitions you should check this file and uncomment > > the line you need most. > >Looking at killpower, I thought this meant I should pick one of > >#mount -n -o ro /usr > >#mount -n -o ro /var > >This is not consistent with the comment in killpower, and, on > >reflection, probably not the intent. > > > >The NEWS file also has a comma splice after "killpower". > >I suggest this revision: > > I have moved powerfail file from /var/spool to /etc. > > Also, there's a new file in /etc/apcupsd which is called > > killpower. If you have /usr/ and /var on different > > partitions you should check this file and uncomment > > the line or lines you need. > > > yep, it's better a dot instead of a comma, however i don't think > it's a good idea to upgrade apcupsd for a missing comma inside NEWS > file. That alone, no. However, the message about "the line you need most" might lead someone to uncomment only one of the lines. > > > >Oops, missed one other problem in the writeup. There's an "and" that > >should be an "or," I believe. So.... > > I have moved powerfail file from /var/spool to /etc. > > Also, there's a new file in /etc/apcupsd which is called > > killpower. If you have /usr/ or /var is not on the root > > partition you should check killpower and uncomment > > the line or lines you need. > uhm.. i'm not a native english speaker however that "and" means > that if u have both partition, or just one of them, u need to uncomment > the file. > In my language this can be done by using "and" , don't know for > >english.
In English, the sense is the same as in mathematics, "A and B" is true only if both A and B are true. Thus, if, for example, /usr is on the root partition while /var is not, the test in "If /usr and /var are not on the root partition" is false, and the rest of the sentence does not apply. I guess the original "If you have /usr and /var on different partitions" actually works, though it's unclear "different" from what. Different from each other? Different from the root partition? Under the former reading (which is the literal meaning of the sentence), the test actually works. Unless /usr and /var are both on / they will be on different partitions (I think--maybe there are some weird cases with the new mount features). > > > Regards > Samuele > As I said, these are all pretty minor issues. They are only significant to the extent they would lead someone astray. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
