Andrey

I've tried to kind of fix the situation by adding more `master` servers to the cluster. `volume` servers aren't dissappearing from `master`'s ui now, but migration process fails. ``` Feb...

@kmlebedev thanks for pointing out, i'll try to add others and write back about the result.

Nope, didnt' helped. After a while still getting the same problem on leader `master` ``` Mar 24 14:45:22 wm-mb-00 weed[25596]: tier move volume 1168182: volume 1168182 not found Mar 24...

Any thoughts? Still experiencing the problem even with latest version.

Hm, seems like this commit is already in 2.98. Gotta check it, because i'm on 2.97 yet. I'll write back if it fixes the problem.

Well, nope, still happening for some reason ``` Apr 21 16:11:02 wm-mb-00 weed[6782]: I0421 16:11:02 6782 master_server.go:277] executing: volume.tier.move [-fullPercent=0 -fromDiskType=ssd -toDiskType=hdd -quietFor=18h -force] Apr 21 16:11:02 wm-mb-00 weed[6782]: tier...

I'm on Debian 10.12. Master processes are started with the following flags: ``` master -raftHashicorp -volumeSizeLimitMB=5000 -peers=192.168.65.93:9333,192.168.65.94:9333,192.168.65.42:9333 -ip.bind=0.0.0.0 -defaultReplication=000 -metricsPort=9327 -mdir=/var/lib/weed/master ```

Could you be a little bit more specific? I have examined my volume servers, the only logs about 8119 are: ``` Jun 15 07:55:49 whv-offline-00 weed[27421]: I0615 07:55:49 27421 store.go:130]...

Yes, you are right, it is wm-offline-02. Anyways, simple `filer` restart fixed the whole situation. How about a retry? It doesn't happen on daily basis, but once in 1-2 weeks...

The main impact of deleting logs is a broken sync between filers. I mean if you use leveldb or other embed storage for filers, you won't be able to start...