[
https://issues.apache.org/jira/browse/HBASE-22301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16826205#comment-16826205
]
Andrew Purtell commented on HBASE-22301:
----------------------------------------
No. The problem is GC activity is indistinguishable from real slow syncs if you
only examine a single data point, unless you set a very high threshold, and
then we would probably not trigger enough to make a difference. Data from our
incident shows a train of slow sync warnings, a few peaks at 1-3 seconds.
Unlikely triggering only on the rare peak outliers would have made a
difference. The conservative 10s trigger in this patch would never have been
reached. Instead, if we triggered on trains of smaller data points in the range
of 200-600ms the mitigation would have fired enough to make a difference and
these trains correlated to real problems not GC activity. And by GC activity I
mean that of the regionserver process. As you probably know any one or a
handful of slow sync warnings can be false positives due to GC rather than real
latency on the pipeline. It makes things difficult here. We can try to avoid
false positives either by setting a high latency threshold or by waiting for an
unusual number to occur within some window of time. There are patches for
review that take either approach. It would seem the high threshold approach may
not offer enough mitigation in practice given the data on hand. At any rate the
thresholds are tunable and can be experimented with in production to find the
right trade off, and the feature is self limiting so slow sync triggered log
rolls do not become a problem themselves.
> Consider rolling the WAL if the HDFS write pipeline is slow
> -----------------------------------------------------------
>
> Key: HBASE-22301
> URL: https://issues.apache.org/jira/browse/HBASE-22301
> Project: HBase
> Issue Type: Improvement
> Components: wal
> Reporter: Andrew Purtell
> Assignee: Andrew Purtell
> Priority: Minor
> Fix For: 3.0.0, 1.5.0, 2.3.0
>
> Attachments: HBASE-22301-branch-1.patch, HBASE-22301-branch-1.patch
>
>
> Consider the case when a subset of the HDFS fleet is unhealthy but suffering
> a gray failure not an outright outage. HDFS operations, notably syncs, are
> abnormally slow on pipelines which include this subset of hosts. If the
> regionserver's WAL is backed by an impacted pipeline, all WAL handlers can be
> consumed waiting for acks from the datanodes in the pipeline (recall that
> some of them are sick). Imagine a write heavy application distributing load
> uniformly over the cluster at a fairly high rate. With the WAL subsystem
> slowed by HDFS level issues, all handlers can be blocked waiting to append to
> the WAL. Once all handlers are blocked, the application will experience
> backpressure. All (HBase) clients eventually have too many outstanding writes
> and block.
> Because the application is distributing writes near uniformly in the
> keyspace, the probability any given service endpoint will dispatch a request
> to an impacted regionserver, even a single regionserver, approaches 1.0. So
> the probability that all service endpoints will be affected approaches 1.0.
> In order to break the logjam, we need to remove the slow datanodes. Although
> there is HDFS level monitoring, mechanisms, and procedures for this, we
> should also attempt to take mitigating action at the HBase layer as soon as
> we find ourselves in trouble. It would be enough to remove the affected
> datanodes from the writer pipelines. A super simple strategy that can be
> effective is described below:
> This is with branch-1 code. I think branch-2's async WAL can mitigate but
> still can be susceptible. branch-2 sync WAL is susceptible.
> We already roll the WAL writer if the pipeline suffers the failure of a
> datanode and the replication factor on the pipeline is too low. We should
> also consider how much time it took for the write pipeline to complete a sync
> the last time we measured it, or the max over the interval from now to the
> last time we checked. If the sync time exceeds a configured threshold, roll
> the log writer then too. Fortunately we don't need to know which datanode is
> making the WAL write pipeline slow, only that syncs on the pipeline are too
> slow and exceeding a threshold. This is enough information to know when to
> roll it. Once we roll it, we will get three new randomly selected datanodes.
> On most clusters the probability the new pipeline includes the slow datanode
> will be low. (And if for some reason it does end up with a problematic
> datanode again, we roll again.)
> This is not a silver bullet but this can be a reasonably effective mitigation.
> Provide a metric for tracking when log roll is requested (and for what
> reason).
> Emit a log line at log roll time that includes datanode pipeline details for
> further debugging and analysis, similar to the existing slow FSHLog sync log
> line.
> If we roll too many times within a short interval of time this probably means
> there is a widespread problem with the fleet and so our mitigation is not
> helping and may be exacerbating those problems or operator difficulties.
> Ensure log roll requests triggered by this new feature happen infrequently
> enough to not cause difficulties under either normal or abnormal conditions.
> A very simple strategy that could work well under both normal and abnormal
> conditions is to define a fairly lengthy interval, default 5 minutes, and
> then insure we do not roll more than once during this interval for this
> reason.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
