I really hate for my first post in the forums to be this, but here we go...
We upgraded our servers to 5.0.3 last night. We have three servers, two CentoOS 5 and one Ubuntu 6.06. One of the CentOS servers and the Ubuntu servers are reporting that partitions are almost full when they are actually well below 10% full. We get errors in /var/log/messages every 10 minutes as well as an email warning us. What was really interesting was that we did actually have a partition that got 97% full (we were using it to dump files temporarily and it didn't get cleaned up) and it wasn't reported to to the logs. I cleaned up the partition, and all is well there but it wasn't reported correctly in the logs either.
Here's what we have been seeing on the CentOS box:
/dev/sda1 is mounted as /boot and given just under 3GB. When it was 4% full it was reported as 91% full by zimbramon. I created a 1GB file of zeros using dd and it was 38% full and reporting correctly in the zimbra logs. Then, I overwrote the file with a 2.5GB file to push it up to 90% and it doesn't warn us with an email that it's almost full.
/dev/sda2 is mounted as /var and is given 95G and was 4% full as well and being reported by zimbramon as 91% full again. I added a 6GB file the same way so it was 10% full and it was being reported in the logs as 85% used. Then I overwrote that file with an 8GB file to push it to 12% full and it's now being reported by zimbramon as being 83% full.
Right now I'm assuming the same will happen with our Ubuntu box as well.
Is anyone else running 5.0.3 yet? I just saw the post about 5.0.3 being pulled so perhaps we're the only ones with this issue so far?
Sorry if I rambled, I tried to give any information I had that would be helpful.