On Jul 03 11:26:44, Amit Kulkarni wrote: > >> > Recently, processes started to die for reasons unknown, as in > >> > > >> > pid 20260 (postgres): user write of 118784@0x28052000 at 159088 failed: > >> > 14 > >> > pid 1872 (cron): user write of 118784@0x2b1e3000 at 30224 failed: 14 > >> > > >> > 14 is EFAULT as per sys/errno,h > >> > - what can be causing it? > >> > >> first check if filesystem is full. > > > > Filesystem Size Used Avail Capacity Mounted on > > /dev/wd0a 196M 113M 73.6M 61% / > > /dev/wd0d 490M 421M 44.2M 90% /usr > > /dev/wd0e 1.6G 382M 1.1G 25% /usr/local > > /dev/wd0g 48.5M 7.4M 38.7M 16% /var > > /dev/wd0h 49.0M 22.8M 23.7M 49% /var/log > > /dev/wd0i 49.0M 20.2M 26.4M 43% /var/mail > > /dev/wd0j 197M 23.6M 164M 13% /var/mysql > > /dev/wd0k 97.8M 39.9M 53.0M 43% /var/postgresql > > /dev/wd0l 49.0M 16.3M 30.3M 35% /var/spool > > /dev/wd0m 4.9G 3.3G 1.3G 71% /var/www > > /dev/wd0n 1005M 34.0K 954M 0% /tmp > > /dev/wd0o 5.0G 2.9G 1.8G 61% /home > > > > Can you please elaborate how a lack of disk space > > would result in an EFAULT? > > write failed, one problem could be disk is full.
Well, apparently that is not the case: there is space on disk. Anyway, failing to write to disk is not an EFAULT, is it? > >> how much is swap configured? > > > > swapctl: no swap devices configured > > > > The machine doesn't do much, and so far the 128M of RAM was enough. > > I will reinstall with a swap partition if it makes a difference - > > should lack of swap result in an EFAULT? > > with low memory machines, you want to load it, and yet don't allow > memory to grow with no swap and datasize-cur limits? some of the > processes will fail eventually. THe machine is mostly idle. It serves as a gateway, pf, dhcp server, resolver and mailserver for the internal "network" of a few laptops, and a webserver for about three zero-traffic sites. 128M is not much of course, but the machine is far from "loaded". > this is another problem. configure > swap to be double the RAM size, or if you insist on loading the > machine about 4 times the RAM size. you obviously forgot to or didn't > want to configure swap. is the "twice the RAM" mantra still valid today? > >> > It usually happens under a stress (the machine is pretty scarse > >> > on resources, so 'stress' can be accepting a batch of DNS queries). > >> > For example, the postgres EFAULT above happened exactly when a batch > >> > of emails arrived. > >> > > >> > Is this a memory problem? My login.conf limits all classes > >> > to datasize-cur=100M. Can this be a memory HW problem?
