Closed sokada1221 closed 5 years ago
@shinnosuke-okada what version of the pump?
@xiaojingchen
image: pingcap/tidb-binlog:v2.1.14
@shinnosuke-okada the new version pump config have a parameter stop-write-at-available-space
,
# stop write when disk available space less then the configured size
# 42 MB -> 42000000, 42 mib -> 44040192
# default: 10 gib
# stop-write-at-available-space = "10 gib"
so if the disk space is less than 10G, the pump will stop writing. you need to use a larger disk or change the parameter. But current operator can not support to change this parameter directly.
In production environments, it is recommended to allocate larger disks for Pump; In the other hand, we will improve Operator's configuration modification to make it more flexible and easy to use. this is a relate PR: #638
@xiaojingchen Thanks a lot. These are very helpful! Let me see if I can unblock myself based on your information.
@xiaojingchen I worked around by allocating more disk space for pump services. I think we have 2 issues here:
tidb-cluster
's default pump config shouldn't come with 10GiBpump
should print a better error message for the case when stop-write-at-available-space
is causing the no available space
errorI'll make a PR for issue 1 immediately, and consider that as a solution to this issue. Planning to increase default disk size for pumps from 10GiB to 20GiB.
@shinnosuke-okada thanks for your feedback, that is very useful to us, and please create an new issue if it's convenient for you.
Logged https://github.com/pingcap/tidb-binlog/issues/679 to address the error message issue. As the problem on TiDB-operator side is addressed, closing this ticket.
Bug Report
What version of Kubernetes are you using?
What version of TiDB Operator are you using?
What storage classes exist in the Kubernetes cluster and what are used for PD/TiKV pods?
What's the status of the TiDB cluster pods?
What did you do?
What did you expect to see? Successful deployment with correctly functioning pump
What did you see instead? Pump log is filled with these error messages:
Eventually, it starts printing the following info messages:
TiDB log is filled with these warning messages:
Eventually TiDB stops responding as it fails to communicate with pump:
But pump has 10GiB of space configured and allocated. Disk space availability is correctly reported on Grafana, and it can also be confirmed inside the pod: