Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
Exactly. But as I was looking it up, I was surprised to find the warning in the sync(2) manpage that, although modern Linuxes now wait till the write is complete before the system call returns, it's still not safe to power down immediately because the data may be in the drive's on-board cache and not yet on the drive's platters. Mark On 6/16/2012 2:03 PM, Jerry Feldman wrote: > On 06/16/2012 01:23 PM, Jack Coats wrote: >> It may have been folk-lore, but the 'standard procedure' where I >> worked, before shutting down a server on purpose, was, from root, >> issue 3 commands separately when we wanted an orderly shutdown, but it >> was 'urgent'. >> >> sync >> sync >> halt >> >> I am sure there is some basis in history that had a basis in fact at >> the time. Even if the 'fact' was based on 'observation' rather than >> reality. >> > See my other reply. With the old file systems, flushing and saving the > buffers was important, The halt(1) command was an immediate halt where > the Shutdown command was a script that did a sync. Generally with the > modern fast CPUs, kernels, and file systems, I don't think a sync will > buy you anything, but it will not cost you. In the olden days when you > used to do useful things, Unix systems were single CPU, and file systems > were slow. I think you remember the stone slab hard drives. What you did > not want it to halt in a partial write, then get into fsck hell. > > > > _______________________________________________ > Discuss mailing list > Discuss at blu.org > http://lists.blu.org/mailman/listinfo/discuss
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |