Feature development is complete and I have pushed code to github, but I still need to add some test cases.
YuanSheng Wang <membp...@apache.org> 于2022年2月9日周三 15:26写道： > the read-only way is the right way. > > this is a draft PR: https://github.com/apache/apisix/pull/6202 > > zaunist is working on it. > > After this PR is merged, APISIX does not need to write etcd most of the > time. > > welcome Jie123108 to join this PR, make it faster ^_^ > > > On Sun, Feb 6, 2022 at 10:15 PM Ming Wen <wenm...@apache.org> wrote: > > > I agree with you, Apache APISIX's DP is read-only for etcd, and as long > as > > etcd can be read, it should work fine. > > > > Thanks, > > Ming Wen, Apache APISIX PMC Chair > > Twitter: _WenMing > > > > > > jie123108 <jie123...@163.com> 于2022年2月6日周日 22:05写道： > > > > > Hi > > > > > > > > > The etcd problem, I know, can be solved by turning on compact option. > > > My question is, why is it that if etcd is not writable, apisix won't > > start > > > and can't be served properly? > > > For someone who is experiencing this problem for the first time, it may > > > take a few minutes or more to resolve the etcd issue, and it is not > > > acceptable for apisix to be out of service during this time > > > > > > > > > > > > > > > Original Message > > > Sender: Zhiyuan Ju<juzhiy...@apache.org> > > > Recipient: dev<dev@apisix.apache.org> > > > Date: Sunday, Feb 6, 2022 21:41 > > > Subject: Re: when etcd database space exceeded. apisix does not start > > > properly > > > > > > > > > Hi Jie123108, Please have a try according to > > > https://github.com/etcd-io/etcd/issues/11947 , and I noticed Bilibli > has > > > the same problem, see > > > https://xie.infoq.cn/article/95c80dcb18ebc2c26237868f3 Best Regards! @ > > > Zhiyuan Ju <">https://github.com/juzhiyuan> jie123108 < > jie123...@163.com > > > > > > 于2022年2月6日周日 21:13写道： > Hi > Happy Chinese New Year! > When the etcd > > > database space is exceeded, apisix is now unable to start, > and > > therefore > > > unable to provide services. Is this reasonable? > > > The error message > > at > > > startup is as follows: > >/usr/local/openresty/luajit/bin/luajit > > > ./apisix/cli/apisix.lua init > >/usr/local/openresty/luajit/bin/luajit > > > ./apisix/cli/apisix.lua init_etcd > >got malformed key-put message: > > > "{"error":"etcdserver: mvcc: database > space > > > exceeded","message":"etcdserver: mvcc: database space > > > > exceeded","code":8}" from etcd "http://etcd:2379/v3/kv/put"; > > > etcd > > is > > > a configuration center, and when its space is exceeded, it can > still > > > provide basic read capabilities. I think the services that depend on > > it > > > should be able to provide basic read capabilities as well. An error > > > > should only be reported when the service is written > > > > > -- > > *MembPhis* > My GitHub: https://github.com/membphis > Apache APISIX: https://github.com/apache/apisix >
