RealisticBiomes loses connection to database

[ghost]

This has been needed for some time. If we are going to do sharding, then we will need to get this fixed anyway so consider it normal priority as part of sharding development.

[ttk2]

agreed. Also we might want to consider moving it to Reddis, probably not
worth it currently, but if there are issues with the ram cache, Reddis is a
keystore, so just use the chunk chords as the key and plants as the
content, since reddis ram caches it will handle all of that, with no other
dependencies it would be an easy conversion and might some day be easier
than improving the ram cache if needed.

On Sat Feb 14 2015 at 6:36:37 PM Dr. Andrew Jawa notifications@github.com
wrote:

This has been needed for some time. If we are going to do sharding, then
we will need to get this out of the way anyway so consider normal priority
as part of sharding development.

—
Reply to this email directly or view it on GitHub
#10.

[erocs]

You'd still have the latency of talking to the reddis server vs. in-proc memory access. You would also need to setup reddis to persist this storage. And that's another potential data store you need to backup and maintain.

[ttk2]

good points, the additional latency is probably reason enough alone. Reddis
does not persist by default?

On Sat Feb 14 2015 at 6:44:35 PM erocs notifications@github.com wrote:

You'd still have the latency of talking to the reddis server vs. in-proc
memory access. You would also need to setup reddis to persist this storage.
And that's another potential data store you need to backup and maintain.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[erocs]

It does... every 5 minutes by default with RDB. You'd want to enable both persistence layers, AOF and RDB, since AOF gives the ability to save every second. Then you'd backup the RDB unless there was a reddis/VM crash and the AOF needed to be used.

[ttk2]

every 5 minutes sounds pretty ok, one second saving might be a bit much.

On Sat Feb 14 2015 at 6:57:20 PM erocs notifications@github.com wrote:

It does... every 5 minutes by default with RDB. You'd want to enable both
persistence layers, AOF and RDB, since AOF gives the ability to save every
second. Then you'd backup the RDB unless there was a reddis/VM crash and
the AOF needed to be used.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[plebes]

Is this still a bug? If so, is it for certain a problem with the RB plugin and not the database or other part of the system?

[ttk2]

We have not gotten it in a while. I think it's fixed. Yes we know it was RB
connections expire and its the plugins job go reopen them if they do.

On Sat, Jun 13, 2015, 3:05 PM plebes notifications@github.com wrote:

Is this still a bug? If so, is it for certain a problem with the RB plugin
and not the database or other part of the system?

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[erocs]

It should be the connection pool's job to re-open the connection. RB
doesn't use one though.

On Sat, Jun 13, 2015 at 1:12 PM, ttk2 notifications@github.com wrote:

We have not gotten it in a while. I think it's fixed. Yes we know it was RB
connections expire and its the plugins job go reopen them if they do.

On Sat, Jun 13, 2015, 3:05 PM plebes notifications@github.com wrote:

Is this still a bug? If so, is it for certain a problem with the RB
plugin
and not the database or other part of the system?

—
Reply to this email directly or view it on GitHub
<
#10 (comment)

.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[ProgrammerDan]

Yeah I was ranting to Gipsy about the RB database code a few hours back.
It's pretty bad, but if it's working semi-stable then not priority atm.

On Sat, Jun 13, 2015 at 4:42 PM, erocs notifications@github.com wrote:

It should be the connection pool's job to re-open the connection. RB
doesn't use one though.

On Sat, Jun 13, 2015 at 1:12 PM, ttk2 notifications@github.com wrote:

We have not gotten it in a while. I think it's fixed. Yes we know it was
RB
connections expire and its the plugins job go reopen them if they do.

On Sat, Jun 13, 2015, 3:05 PM plebes notifications@github.com wrote:

Is this still a bug? If so, is it for certain a problem with the RB
plugin
and not the database or other part of the system?

—
Reply to this email directly or view it on GitHub
<

#10 (comment)

.

—
Reply to this email directly or view it on GitHub
<
#10 (comment)

.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[ttk2]

An open source connection pool setup would probably be nice across all
plugins.

On Sat, Jun 13, 2015, 10:04 PM Daniel Boston notifications@github.com
wrote:

Yeah I was ranting to Gipsy about the RB database code a few hours back.
It's pretty bad, but if it's working semi-stable then not priority atm.

On Sat, Jun 13, 2015 at 4:42 PM, erocs notifications@github.com wrote:

It should be the connection pool's job to re-open the connection. RB
doesn't use one though.

On Sat, Jun 13, 2015 at 1:12 PM, ttk2 notifications@github.com wrote:

We have not gotten it in a while. I think it's fixed. Yes we know it
was
RB
connections expire and its the plugins job go reopen them if they do.

On Sat, Jun 13, 2015, 3:05 PM plebes notifications@github.com wrote:

Is this still a bug? If so, is it for certain a problem with the RB
plugin
and not the database or other part of the system?

—
Reply to this email directly or view it on GitHub
<

#10 (comment)

.

—
Reply to this email directly or view it on GitHub
<

#10 (comment)

.

—
Reply to this email directly or view it on GitHub
<
#10 (comment)

.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[plebes]

Suggestions? BoneCP?

[erocs]

I was originally thinking C3P0 but that looks like it would work too.

On Wed, Jun 17, 2015 at 10:18 PM, plebes notifications@github.com wrote:

Suggestions? BoneCP?

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

[ttk2]

just pick one that works and run with it.

[rourke750]

Just put in in CivModCore since thats what its used for...

On Thu, Jun 18, 2015 at 7:56 AM, ttk2 notifications@github.com wrote:

just pick one that works and run with it.

—
Reply to this email directly or view it on GitHub
#10 (comment)
.