Aaron Kulkis wrote: > Ken Schneider wrote: >> Aaron Kulkis wrote: >>> Ken Schneider wrote: >>>> Aaron Kulkis wrote: >>>>> Clayton wrote: >>>>>>> Problem solved: I filled the partition with /tmp on it, so nothing >>>>>>> could be parked there. I found out when I tried running the sax2 man >>>>>>> page which crashed but gave me the info I needed >>>>>> Interesting. I hadn't considered that a full /tmp partition would do >>>>>> it. I don't have it on a separate partition on any of the systems I >>>>>> run, so this never came up. I'll be tucking that bit of info away >>>>>> for >>>>>> future use. >>>>> /tmp can fill up even if it's on the root partition...of course, >>>>> if that happens, then your root partition is full, too. >>>>> >>>>> Personally, I don't like ANY unnecessary file I/O on my root >>>>> partition, so /tmp always gets its own partition >>>> I/O is still I/O that has to be handled by the _disk_ not the >>>> _partition_, so it matters not if /tmp is on it's own partition unless >>>> it is also on a different disk. >>> Really? >>> >>> So if the /tmp directory is corrupted, and it is on its >>> own partition (and therefore, a separate filesystem), >>> this corrupts the root filesystem how, exactly? >>> >> >> You weren't talking about corruption but only disk I/O. > > A filesystem can't get corrupted if there's no write > activity on it.
DUH! > >> Disk corruption is another matter. > > No, they're directly related. > >> Disk I/O affects all partitions on the disk not just one. > > the overall wear of read/write head and it's controller arm > is effected, along with the data on the particular > partition/filesystem being written to. Which affect the other partitions because they _cannot_ be written to at the same time, duh! > > Remember, it's a physical device doing a physical > activity...MODIFYING the data at a specific location > on the physical platter(s). > >> If the disk is writing to a /tmp partition it _is_ going to affect >> the / partition due to head movement etc. > > But that corrupts the / partition how, exactly? You are the one that brought up corruption, not me. > > You're not an engineer, are you? > > You don't have to be an engineer to realize you can't perform write operations to two different partitions at the same time on the same disk. I still say I/O is I/O, you're either writing to or reading from a specific place on one platter of the disk which prevents reading/writing from any other place on the disk ( I/O ). Ken -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]