Hi Alan,
    Thank you for your quick response! It is helpful! I did "udevadm trigger 
--subsystem-match=block” using root, which seems doesn’t work, so I did it 
again using user ubuntu, now it is working! Thank you!

But I have another question, is there anyway to have ATS store content in the 
RAM for a while before writing to disk?
Because my origin sends some content, then there is some chance this content 
may be changed within a few seconds (origin will push again), if I can store 
the content in the RAM for while, it can potentially save some disk 
reads/writes.


Best,
Jason


On Mar 4, 2019, 11:53 -0500, Alan Carroll <[email protected]>, 
wrote:
> 1) Did you do an `ls -l /dev/nvme*` to verify the permissions are what you 
> think you set in the udev file?
> 2) Does group "ubuntu" have write permission?
> 3) Did you set the process group for `traffic_server` to be "ubuntu"? See 
> https://docs.trafficserver.apache.org/en/8.0.x/admin-guide/files/records.config.en.html#proxy-config-admin-user-id
>  - the primary group of the user ID will be used as the effective group ID of 
> `traffic_server`.
>
> > On Mon, Mar 4, 2019 at 9:59 AM Jason Yang <[email protected]> 
> > wrote:
> > > Hi
> > > I am trying to setup ATS 8.0.2 on AWS EC2 using m5d instance (with local 
> > > disk)
> > >
> > > I am running traffic server using user ubuntu and the instance has disk 
> > > like following
> > > nvme0n1 259:0 0 279.4G 0 disk
> > > nvme1n1 259:1 0 279.4G 0 disk
> > > nvme2n1 259:2 0 8G 0 disk
> > > └─nvme2n1p1 259:3 0 8G 0 part /
> > >
> > > I added the following lines to /etc/udev/rules.d/51-cache-disk.rules
> > >  SUBSYSTEM==block, KERNEL==nvme[0123456789], GROUP:=ubuntu
> > >  SUBSYSTEM==block, KERNEL==nvme[0123456789]n1, GROUP:=ubuntu
> > >
> > > and my storage.config
> > >  /dev/nvme0n1 label=cache.disk.0
> > >  /dev/nvme1n1 label=cache.disk.1
> > >
> > > But it seems it is not working and I am getting "WARNING: unable to open 
> > > '/dev/nvme0n1': Permission denied”.
> > >
> > >
> > > Am I doing something wrong? Thank you!
> > >
> > >
> > > Jason
> > >
> > >
> > >
> > >

Reply via email to