-
Notifications
You must be signed in to change notification settings - Fork 1.1k
-
I saw a similar discussion in the repository, but it looks like that was because of no space left in the volume.
I am running CouchDB 3.2.2 with docker image and I have a couple of databases being replicated. Apparently for the last month or so, database replications are fairly unstable and I see many crashes like:
[error] 2024年06月21日T17:27:31.189103Z couchdb@127.0.0.1 <0.12603.0> -------- gen_server <0.12603.0> terminated with reason: no match of right hand value {error,{options,{socket_options,[{nodelay,true},{keepalive,true}],einval}}} at ibrowse_http_client:send_req_1/8(line:981) <= gen_server:try_handle_call/4(line:661) <= gen_server:handle_msg/6(line:690) <= proc_lib:init_p_do_apply/3(line:249)
on the node that creates the replication documents.
This is preceded by a stacktrace and timeout replication errors, e.g.:
[error] 2024年06月21日T17:27:31.184332Z couchdb@127.0.0.1 <0.14490.0> -------- Replicator, request GET to "https://***.cern.ch/couchdb/workqueue/9c2ce35b073c5f9c4cfda28a196e5cfb?revs=true&open_revs=%5B%2216
7-7ed0af9669fe0843ba85d9c3a2cba347%22%5D&latest=true" failed due to error timeout
but this timeout cannot be seen in the other node and Apache frontend either. On the other side, the request went through without any issues:
[notice] 2024年06月21日T17:12:30.617564Z nonode@nohost <0.17336.7498> 2ae6533081 ***.cern.ch:5984 ip_adress undefined GET /workqueue/9c2ce35b073c5f9c4cfda28a196e5cfb?revs=true&open_revs=%5B%22167-7ed0af9669fe0843ba85d9c3a2cba347%22%5D&latest=true 200 ok 3
Does anyone understand what the problem could be and/or how I could investigate these instabilities?
Beta Was this translation helpful? Give feedback.
All reactions
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment