I run a similar script in R, but it seems that OK, setting prior of the b

How is your prior of b setting?

Has your MCMC chain already converged?

PS: I tried something as below as you set b as fixed. But I also tried some other prior of b, it is OK.

N <- 100
x <- rgamma(N, 5, 1)


write("
model {
    for (i in 1:length(x)) {
        x[i] ~ dgamma(a, b)
    }
    a ~ dgamma(3, 1)
    b <- 1
}",
file="example1.bug")

jags <- jags.model('example1.bug',
                   data = list('x' = x),
                   n.chains = 4,
                   n.adapt = 100)


mcmc.samples <- coda.samples(jags,
             c('a', "b"),
             n.iter = 10000)

summary(mcmc.samples)