I am not the one that is qualified to answer but I guess it depends on how much you transact. @UB95JRKM3 @j.massa @U1C36HC6N

How do I force it to persist? So I can point traffic to the new instance

The data is indeed persisted to the konserve backend on every transaction, so there shouldn't be a problem moving the data with the file backend. I can't say anything about other backends though or specifics of datahike-server.

huh :thinking_face: I guess then all backends should do the same. at least from what I know.

well, other external backends are basically services themselves, so it depends on when they persist their data to disk

right

As long as datahike dispatches to on each transaction I'm good

I know konserve-jdbc also dispatches the write immediately

Then you should be good without any further action

I checked last night. I'm on 0.1510. It's definitely not syncing. I connected 3 datahike-server instances. And they didn't have the same data.

It's super worry because now I don't know if I reboot if I'll loose data

What do you mean with 0.1510? And why three datahike-server instances?

I'm on datahike 0.1510. I connected 3 different datahike-servers to the same mysql db.

0.6.1510?

Ah sorry missed a number 0.5.1510

The running instance has the most up to date data. The other two are quite far behind.

Datahike did not update the pointer when transacting data on another instance, you need https://github.com/replikativ/datahike/releases/tag/0.6.1539

But if I reboot my server there won't be any data loss right?

all the transacted data should be there

but keep in mind when reading and transacting from different server instances you have inconsistencies

I don't plan on doing both at the same time. It's just when I want to switch providers. I can power down one and move to another.

the pipeline has an error

lms

clojars is down today

released it now 0.1.2

Thanks. I'm trying to update datahike-jdbc but I'm getting a whole bunch of error. I'm not too familar with deps.edn but my guess is that the pom isn't reflecting what's in the deps.edn

try clj -X:deps mvn-pom

If I don't add

com.github.seancorfield/next.jdbc {:mvn/version "1.3.874"}
        com.mchange/c3p0 {:mvn/version"0.9.5.5"}

are you using the development branch?

Tried both. Is my understanding that the pom.xml should match deps.edn correct?

in konserve-jdbc

the pom is generated with the command above

What I mean is: That command generates a pom.xml. Presumably it runs before deploying to clojars.

The current pom.xml doesn't match deps.edn in the repo. Which means clojars doesn't have the required deps either

• konserve on clojars is 0.6.0. In deps.edn 0.7.285

Should I run that command and do a pull request

when you look at the circleci/config you can see that first the pom is updated and then the jar is built. To avoid confusion the pom should not have dependencies in it. somehow it got committed. but from what I see the jar is always built with updated pom.xml even if not committed to the git repo.

I assume that clojars is pulling the deps from the repo not from the built jar. don't know why though

you want to do a short call?

In 14 min?

ok, send me a link

https://meet.google.com/vtd-neta-tqc

looks better now https://clojars.org/io.replikativ/konserve-jdbc

Awesome 🎉