[
https://issues.apache.org/jira/browse/ACCUMULO-1585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13713364#comment-13713364
]
Basit Mustafa commented on ACCUMULO-1585:
-----------------------------------------
A very perfunctory/basic examination of what's going on at startup has me
believing that implementing a basic (non-config flag based optional/conditional
behavior) fix would go something like this:
In org.apache.accumulo.server.tabletserver.TabletServer:
1) Add ivar to store the hostname String exactly as passed to the config(String
hostname) method (from looking at the output of this method's first log
statement, it appears it not yet resolved, but as typed in config, this is a
good thing.
2) From here, a few possible paths are possible:
A. One COULD just say let's modify getClientAddressString() to not return a
resolved address. That is assuming this method's contract does not guarantee an
IP:PORT String and that all callers are safe using an FQDN or whatever the
config file had verbatim. The documentation/comment does not have a specific
contract, but the lack of strong typing of the return value to an IP:PORT type
(e.g. INetSocketAddress or something) makes me hopeful this would work
(although could see this blowing up in all kinds of ways, too, if this String
return value is expected to be IP:PORT by callers to getClientAddressString()).
B. If this doesn't work or we know we don't want to go off changing the nature
of this method because it'd violate its unwritten contract/caller expectation
that it return IP:PORT, we could go off and say that we'll only write
FQDN/hostname as passed verbatim into config(String hostname) (now stored in
ivar from #1) in ZooKeeper and keep all Accumulo internals as-is (this works,
IMHO since the internals past this point are all in the same JVM as long as we
write FQDNs to ZK and we won't have the aforementioned schizophrenia because
resolution in the same JVM should be the same barring DNS roundrobining/load
balancing [uh, just don't do this between nodes in an Accumulo cluster :)]).
Then, we're on the hook to go discover where
/accumulo/<instance>/tservers/XXXXXXXXX are read on the client and ensure that
that read does the resolution of the retrieved FQDN/string, or at least just
runs it through AddressUtils.toString().
Obviously, B involves the least changes to Accumulo code as it seems pretty
straightforward since reads/writes to ZK are pretty obvious/unified in a single
set of classes. A is making some large assumptions/leaps about the safety of
changing the format of that String output, I'd feel better about it knowing
what the author of it (and its callers) intended. I haven't done a "who calls
this" analysis to see, I guess I could smoke test it, too, of course. But, B
just seems like the path of much less resistance assuming we're only reading
the value from ZK in one/a few places.
Thoughts? Opinions? Anyone have any experience/know the code better than me to
help shed light on assumptions or come up with C/D/E/F options that would be
better?
Thanks!
> Provide option for FQDN/verbatim data from config files of servers to be
> stored in ZooKeeper rather than resolved IP
> --------------------------------------------------------------------------------------------------------------------
>
> Key: ACCUMULO-1585
> URL: https://issues.apache.org/jira/browse/ACCUMULO-1585
> Project: Accumulo
> Issue Type: Improvement
> Components: tserver
> Environment: All
> Reporter: Basit Mustafa
> Assignee: Eric Newton
> Priority: Minor
> Fix For: 1.6.0
>
> Original Estimate: 12h
> Remaining Estimate: 12h
>
> There are some situations (esp in virtualized/cloud environments) where
> "hardwiring" the IP into ZooKeeper can create reachability issues and an FQDN
> (or, better/also, the verbatim string/line from the concerned config file)
> would fix this problem.
> For example, hostname node1.company.com specified in configuration files
> resolves to an Amazon EC2 *internal* IP of 10.2.3.4 (internal on virtualized
> network). Externally (e.g. from your dev machine, your
> offsite/non-VPN/non-VPCed data center, other client machines on different
> networks/clouds), node1.company.com will resolve to a public IP (e.g. Amazon
> Elastic IP, etc) of something more routeable, like 54.55.56.57.
> Accumulo currently stores 10.2.3.4 in ZooKeeper based on this resolution,
> but, if you try to connect to Accumulo from outside these machines/machines
> in the same cloud/vitualized network/non routeable network, and the same FQDN
> (node1.company.com) resolves to the public address now (54.55.56.57), you
> will not be able to connect, because the Accumulo client will have pulled the
> resolved, and from here, unreachable, IP of 10.2.3.4.
> Using the FQDN (or in some other way allowing for client-side name
> resolution/address translation, although this seems kludgy) would fix this
> issue in a relatively standard way. Ideally, this would not incur a
> performance issue beyond the first resolution assuming the TCP/IP stack is
> doing its job and caching stuff effectively (I assume).
> This doesn't really hurt/break things if you give an option in some config,
> and, really, taking the literal from the file allows you to use whatever you
> want, the ultimate in flexibility.
> See discussion
> http://mail-archives.apache.org/mod_mbox/accumulo-user/201307.mbox/%3CCAGFNOZTMVz0R2e0meDj%3DKqPPPJP6f5baaMqh8%3D07V7NZ8vToJg%40mail.gmail.com%3E
> for more details and others having the same issue.
> I will look into creating a patch for this as soon as I have some time to
> find/look at relevant code portions (I need to find where accumulo is making
> these writes to ZK and if the read FQDNs would need any resolution/their use
> further down the line expects strictly IP or is in host or IP safe API calls,
> etc). Any suggestions on where I can begin this are always appreciated.
> Otherwise, I'll try and submit a patch when I can.
> Figure I'd open this issue to at least provide a discussion on what more
> experienced Accumulo devs and users think and what a solution based on the
> style/patterns accepted for Accumulo development/configuration are. I can
> read the guidelines myself, of course, and will, but someone suggested
> opening an issue, so I am...
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
