Anonymous_User Absent Member.
Absent Member.
684 views

Stuck replica

Running edir 886 on oes2sp3. Have a server that has a couple
subordinate replicas that are stuck in a new state.

I'm trying to clean this up so I can put the correct replicas back on
this server, but I cannot seem to get them to go away.

I've tried ndsrepair -P -Ad on it, doesn't seem to help.

If I try to add replicas to this server, but they always stay in a new
state.

If I select this server in C1, I get an error:

The directory service provider used by consoleOne returned an unknown
error.

javax.naming.NameNotFoundException, remaining name 'server.tree'.

Suggetions?

--
Stevo
Labels (1)
0 Likes
18 Replies
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Tue, 22 Nov 2011 18:01:25 +0000, Stevo wrote:

> Suggetions?


I'd start with a health check to determine why the subrefs are stuck in a
new state. That usually is something to do with a communication issue -
which should be resolved before trying to fix the state issue.

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> I'd start with a health check to determine why the subrefs are stuck
> in a new state. That usually is something to do with a communication
> issue - which should be resolved before trying to fix the state issue.


So my response to Jim's comment is...

Time is synched on these boxes, but when I do a synchronization status
(run from oes2sp3 box that is master of all replicas), I get errors:

-761 errors to the server w/ the replica issues
-603 errors to the server that used to be the master (nw65sp8, master
is now oes2sp3 on sles10sp3)
-672 errors to the same server as the -603 errors
-694 error to other nw65sp8 server that has a couple replicas

One thing I did notice is the trouble server (oes2sp3) is on ds version
20602.00 while the other oes servers are 20605.00. I've tried running
the online update for that box, but it will not pull any more updates.
Any way to install that update manually?


--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Stevo sounds like they 'said':

> One thing I did notice is the trouble server (oes2sp3) is on ds
> version 20602.00 while the other oes servers are 20605.00. I've
> tried running the online update for that box, but it will not pull
> any more updates. Any way to install that update manually?


So my response to Stevo's comment is...

Ok, got that issue resolved, all oes servers are on the same ds version
now.

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Tue, 22 Nov 2011 20:59:18 +0000, Stevo wrote:

> Ok, got that issue resolved, all oes servers are on the same ds version
> now.


That's a good starting point. As Andrew said, the subrefs are created
when a replica is added to a server and the partition has child
partitions.

But I'd go back and do a full health check in iMonitor (start with a
server holding a copy of root, go to "Agent Health" and check that, and
then go to the known servers list and follow the 'agent health' link for
each server in the tree).

Let's see what that tells us about the environment.

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> Let's see what that tells us about the environment.
>
> Jim


So my response to Jim's comment is...

Ok, so I started checking. The only server that does not show good or
marginal is the nw65sp8 server that used to be the master. It shows
Warn on Partition/Replication.

When I follow that link, I get good for [root], and warn for 2 of the
other partitions. Only 3 of the 4 partitions are even listed.

If I look at the other nw65sp8 server (which has the same for
replicas), everything shows as good.

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Wed, 23 Nov 2011 15:13:04 +0000, Stevo wrote:

> Jim Henderson sounds like they 'said':
>
>> Let's see what that tells us about the environment.
>>
>> Jim

>
> So my response to Jim's comment is...
>
> Ok, so I started checking. The only server that does not show good or
> marginal is the nw65sp8 server that used to be the master. It shows
> Warn on Partition/Replication.
>
> When I follow that link, I get good for [root], and warn for 2 of the
> other partitions. Only 3 of the 4 partitions are even listed.
>
> If I look at the other nw65sp8 server (which has the same for replicas),
> everything shows as good.


If you continue to drill down, what error or warning do you eventually
get? iMonitor will tell you why it's giving a warning when you get far
enough in.

(For example, my server shows a red bar warning because I only have a
single replica of [Root] - because it's a single server tree. If I drill
into it, it tells me the reason is that I have fewer than 3 replicas).

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> If you continue to drill down, what error or warning do you
> eventually get? iMonitor will tell you why it's giving a warning
> when you get far enough in.


So my response to Jim's comment is...

I guess I'm not sure where one can see that info. I got things pretty
much cleared up, but the original server that started this whole mess
shows one warning on it in imonitor.

From the root server, I went to known servers. Then to this server's
agent health. In there the Partition/Replication shows a warning.

I go into Partition/Replication, and see a warning on the partition at
my O level. From there I'm not quite sure where to go to actually see
the warning/error. You can go into the partition, or the Replica
Synchronization link, or to the Ring itself.

I ended up going into the Ring, there I can see under the Health
Check:Replica section. In there I see quite a few entries where the
last successful sync and/or last attempted sync is over 36,000 hours
ago, and there are a few warnings next to the Flags listings.

To clarify, I got things 'cleared up' by removing all the replicas of
my partitions from all but the master server and another one that did
not have any issues. Then left things sit for several hours, and when
I tried adding replicas of my partitions, they all went to an ON status
pretty quick.


--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Wed, 23 Nov 2011 22:24:33 +0000, Stevo wrote:

> I guess I'm not sure where one can see that info. I got things pretty
> much cleared up, but the original server that started this whole mess
> shows one warning on it in imonitor.
>
> From the root server, I went to known servers. Then to this server's
> agent health. In there the Partition/Replication shows a warning.
>
> I go into Partition/Replication, and see a warning on the partition at
> my O level. From there I'm not quite sure where to go to actually see
> the warning/error. You can go into the partition, or the Replica
> Synchronization link, or to the Ring itself.


If you go to the "Ring" link, that'll give you a ring health check.

> I ended up going into the Ring, there I can see under the Health
> Check:Replica section. In there I see quite a few entries where the
> last successful sync and/or last attempted sync is over 36,000 hours
> ago, and there are a few warnings next to the Flags listings.


What are the listed warnings? If there's more than 1 server in the ring
with a 'real' (non-subref) copy of the partition, not being in sync for
that long will cause you problems and is probably what's holding up the
subrefs being added.

That's the issue to be resolved before going forward.

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> What are the listed warnings? If there's more than 1 server in the
> ring with a 'real' (non-subref) copy of the partition, not being in
> sync for that long will cause you problems and is probably what's
> holding up the subrefs being added.


So my response to Jim's comment is...

Well as of right now there are no sub-refs anywhere in the tree. Like
I said, I cleaned house on almost all replicas of my partitions. It
seems the partition in question now is the one at my O level.

The errors listed are:

Last Successful Sync (on all servers)
Last Attempted Sync (on one oes2 server, call it e88-dns)
Flags (on e88-dns)

Then have some Suspect errors on:

Receive delta (on all but e88-dns)
Send delta (on e88-dns)
Purge time (on e88-dns)
Perishable Data (on e88-dns)

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Wed, 23 Nov 2011 22:51:32 +0000, Stevo wrote:

> Well as of right now there are no sub-refs anywhere in the tree. Like I
> said, I cleaned house on almost all replicas of my partitions. It seems
> the partition in question now is the one at my O level.
>
> The errors listed are:
>
> Last Successful Sync (on all servers)
> Last Attempted Sync (on one oes2 server, call it e88-dns)
> Flags (on e88-dns)
>
> Then have some Suspect errors on:
>
> Receive delta (on all but e88-dns)
> Send delta (on e88-dns)
> Purge time (on e88-dns)
> Perishable Data (on e88-dns)


How many servers in the replica ring, and what does the partition layout
look like?

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> How many servers in the replica ring, and what does the partition
> layout look like?


So my response to Jim's comment is...

Only 5 servers in the tree, all servers have replicas of all partitions.

Partitions on [root], my org level, and the ou=configuration,o=org and
ou=schema,ou=configuration,o=org levels from my dsfw installation.

Should have probably mentioned this is not a production tree, but would
still like to get this straightened out as I need to add an OES2 server
to this tree to simulate migrating file services from netware to linux.

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Jim Henderson sounds like they 'said':

> How many servers in the replica ring, and what does the partition
> layout look like?


So my response to Jim's comment is...

If I look in the replica synchronization portion, one server is causing
issues as it shows -609 errors on items that I cannot find in this tree.

They all show cn of 0_1, in my SLP info, and I cannot find anything w/
that type of name anywhere in my tree, let alone my SLP.

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

Stevo sounds like they 'said':

> If I look in the replica synchronization portion, one server is
> causing issues as it shows -609 errors on items that I cannot find in
> this tree.
>
> They all show cn of 0_1, in my SLP info, and I cannot find anything w/
> that type of name anywhere in my tree, let alone my SLP.


So my response to Stevo's comment is...

So I did a delete unknown leaf objects on the server in question,
errors all went away. If I do an ndsrepair -E on my master server, I
get 0 errors.

--
Stevo
0 Likes
Anonymous_User Absent Member.
Absent Member.

Re: Stuck replica

On Mon, 28 Nov 2011 17:32:12 +0000, Stevo wrote:

> Stevo sounds like they 'said':
>
>> If I look in the replica synchronization portion, one server is causing
>> issues as it shows -609 errors on items that I cannot find in this
>> tree.
>>
>> They all show cn of 0_1, in my SLP info, and I cannot find anything w/
>> that type of name anywhere in my tree, let alone my SLP.

>
> So my response to Stevo's comment is...
>
> So I did a delete unknown leaf objects on the server in question, errors
> all went away. If I do an ndsrepair -E on my master server, I get 0
> errors.


What does iMonitor's health check show? Does it show sync is working
properly now?

Jim



--
Jim Henderson, CNA6, CDE, CNI, LPIC-1, CLA10, CLP10
Novell Knowledge Partner
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.