Yes, one line per LogEnt= ry. A line produced by phosphor-hostlogger in the zip file has a timestamp= =C2=A0+ log message in it. The timestamp is then filled into LogEntries.

>We propose the following= methods to improve it.
>
>=C2=A0 =C2=A0- Method 1: *D-Bus Signal*; ph= osphor-hostlogger implements an interface
>=C2=A0 =C2=A0which contains a signal. The paylo= ad of the signal should contain
>=C2=A0 =C2=A0timestamps and log messages.=C2=A0 Bm= cWeb registers as a listener and once it
>=C2=A0 =C2=A0receives a signal, it popula= tes a new LogEntry. BmcWeb should implement its
>=C2=A0 =C2=A0own configurable ring= buffer to store log entries received from D-Bus.
>=C2=A0 =C2= =A0- Method 2: *File Watcher*; add file watchers in BmcWeb to monitor the>=C2=A0 =C2=A0lo= g files produced by phosphor-hostlogger. This method is similar to method>=C2=A0 = =C2=A01. But persistence is still a necessary condition.
>=C2= =A0 =C2=A0- Method 3: *obmc-console + bmcweb*: install the console collecti= on and
>=C2=A0 = =C2=A0ring buffer parts of phosphor-hostlogger as a library. Use the librar= y
>=C2= =A0 =C2=A0directly in BmcWeb to collect console logs.
>=C2=A0 = =C2=A0- Method 4: *phosphor-hostlogger + journal + rsyslog + bmcweb*: this<= br>>=C2=A0 =C2=A0a= rchitecture is very similar to what the current OpenBMC uses for
= >=C2=A0 =C2=A0redf= ish-event
>=C2=A0 =C2=A0<https://github.com/openbmc/docs/blob/master/archi= tecture/redfish-logging-in-bmcweb.md>.
>=C2=A0 =C2=A0Add a new schema for log entri= es. Publish journal logs in
>=C2=A0 =C2=A0phosphor-hostlogger. Add file watchers in= BmcWeb to monitor the log files
>=C2=A0 =C2=A0produced by rsyslog. rsyslog shoul= d have log rotation enabled. Persistence
>=C2=A0 =C2=A0is still a necessary conditi= on.
= Perhaps a missing option here is a single dbus object per LogEntry?=C2=A0 I= 'm
not necessarily advocating for that - can I assume that was notconsidered because of performance/scaling concerns?

Yes,= one DBus object per LogEntry is an option. But as you said, this option mi= ght not scale. We want to mimic a near real-time console via Redfish and we= care about=C2=A0latency.

Sincerely,

Nan=

On Tue, Mar 9, 2021 at 6:30 AM Brad Bishop <bradleyb@fuzziesquirrel.com> wrote:

Hi Nan

Thanks for this announcement.=C2=A0 I don't have a lot of specific feed= back
for you, but this sort of introduction to new content is super helpful
for the overall health of OpenBMC.

On Mon, Mar 08, 2021 at 01:45:00PM -0800, Nan Zhou wrote:

>Implementation for Phase 1 is in
>https://gerrit.openbmc-project.xyz/c/= openbmc/bmcweb/+/39093. It is
>based on obmc-console, phosphor-hostlogger, and bmcweb. The basic idea = is

I like how you are making use of these existing projects.

>that phosphor-hostlogger collects host serial console logs via obmc-con= sole
>and generates tarballs in a rotation manner. These tarballs are then >consumed and exposed by a new node in bmcweb log service.

Is it one console line per LogEntry?

>We propose the following methods to improve it.
>
>=C2=A0 =C2=A0- Method 1: *D-Bus Signal*; phosphor-hostlogger implements= an interface
>=C2=A0 =C2=A0which contains a signal. The payload of the signal should = contain
>=C2=A0 =C2=A0timestamps and log messages.=C2=A0 BmcWeb registers as a l= istener and once it
>=C2=A0 =C2=A0receives a signal, it populates a new LogEntry. BmcWeb sho= uld implement its
>=C2=A0 =C2=A0own configurable ring buffer to store log entries received= from D-Bus.
>=C2=A0 =C2=A0- Method 2: *File Watcher*; add file watchers in BmcWeb to= monitor the
>=C2=A0 =C2=A0log files produced by phosphor-hostlogger. This method is = similar to method
>=C2=A0 =C2=A01. But persistence is still a necessary condition.
>=C2=A0 =C2=A0- Method 3: *obmc-console + bmcweb*: install the console c= ollection and
>=C2=A0 =C2=A0ring buffer parts of phosphor-hostlogger as a library. Use= the library
>=C2=A0 =C2=A0directly in BmcWeb to collect console logs.
>=C2=A0 =C2=A0- Method 4: *phosphor-hostlogger + journal + rsyslog + bmc= web*: this
>=C2=A0 =C2=A0architecture is very similar to what the current OpenBMC u= ses for
>=C2=A0 =C2=A0redfish-event
>=C2=A0 =C2=A0<https://github.com/openbmc/docs/blob/master/architecture/redfish-logg= ing-in-bmcweb.md>.
>=C2=A0 =C2=A0Add a new schema for log entries. Publish journal logs in<= br> >=C2=A0 =C2=A0phosphor-hostlogger. Add file watchers in BmcWeb to monito= r the log files
>=C2=A0 =C2=A0produced by rsyslog. rsyslog should have log rotation enab= led. Persistence
>=C2=A0 =C2=A0is still a necessary condition.

Perhaps a missing option here is a single dbus object per LogEntry?=C2=A0 I= 'm
not necessarily advocating for that - can I assume that was not
considered because of performance/scaling concerns?

-brad

--0000000000007e6df805bd34db9a--