Send netdisco-users mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.sourceforge.net/lists/listinfo/netdisco-users
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of netdisco-users digest..."
Today's Topics:
1. Re: Search failed! Please contact your site administrator
(netdisco2) (Oliver Gorwits)
2. Re: Search failed! Please contact your site administrator
(netdisco2) (Oliver Gorwits)
3. Re: Best way to discover Cisco Nexus N5K (Sebastian R?sch)
4. Re: Search failed! Please contact your site administrator
(netdisco2) (Christian Ramseyer)
5. Re: The connected device information is incorrect on
Netdisco. (Chenjie (Jacy, Enterprise Romania GTAC))
--- Begin Message ---
Hi Dan,
Thanks for getting in touch and asking.
On 2015-12-10 19:57, Daniel Bird wrote:
Again, this a 6509, with a fair few modules and ports (382).
The only time I can get it to display any port information is if I
remove almost all the display items, and especially "Connected
Nodes".
I regret that, to my knowledge, this is probably "known and expected
behaviour" with the current design of Netdisco.
The Device Ports page retrieves almost all of the displayed information
for all ports in one go. Paging is synthetic within the browser.
That means the DB query response can balloon if you have a lot of nodes
on one device, and that's certainly the case for the large 6500s (I ran
a few myself in the past and I know this issue well).
One day... I guess we hope to re-engineer the web frontend to retrieve
slices of results as you page, or something similar, but it hasn't been
implemented.
Just in case you're proxying behind apache, you could try increasing
the timeout value in Apache's config. Alternatively, throw resources at
your PostgreSQL server - it will eat as much CPU, RAM, and fast disk as
you can provide. Also run the pgtune script as per the ND2 install docs.
regards,
oliver.
I see there was a mention of it being a memory consumption issue, and
it being looked in to. I wondered if there any progress on this?
We're running the latest version
[netdisco2@netmanage ~]$ ~/bin/localenv cpanm --notest App::Netdisco
App::Netdisco is up to date. (2.033004)
And all the other devices we have seem to list their ports fine,
although we don't have any other kit with anywhere near as many ports
as the 6509
Cheers
Dan
------------------------------------------------------------------------------
_______________________________________________
Netdisco mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
--- End Message ---
--- Begin Message ---
... I should add that there is no reason we could not improve the
performance, but for a lack of time. I agree that it can and should be
better!
So if anyone would like to take a look at the DB queries, or submit
patches for the web frontend, I would be very welcome indeed.
To enable debugging to see the DB queries, run this in your terminal
and visit Netdisco on port 5001:
DBIC_TRACE=1 ~/bin/localenv plackup -p 5001 bin/netdisco-web-fg
regards,
oliver.
On 2015-12-13 21:28, Oliver Gorwits wrote:
Hi Dan,
Thanks for getting in touch and asking.
On 2015-12-10 19:57, Daniel Bird wrote:
Again, this a 6509, with a fair few modules and ports (382).
The only time I can get it to display any port information is if I
remove almost all the display items, and especially "Connected
Nodes".
I regret that, to my knowledge, this is probably "known and expected
behaviour" with the current design of Netdisco.
The Device Ports page retrieves almost all of the displayed
information
for all ports in one go. Paging is synthetic within the browser.
That means the DB query response can balloon if you have a lot of
nodes
on one device, and that's certainly the case for the large 6500s (I
ran
a few myself in the past and I know this issue well).
One day... I guess we hope to re-engineer the web frontend to
retrieve
slices of results as you page, or something similar, but it hasn't
been
implemented.
Just in case you're proxying behind apache, you could try increasing
the timeout value in Apache's config. Alternatively, throw resources
at
your PostgreSQL server - it will eat as much CPU, RAM, and fast disk
as
you can provide. Also run the pgtune script as per the ND2 install
docs.
regards,
oliver.
I see there was a mention of it being a memory consumption issue,
and
it being looked in to. I wondered if there any progress on this?
We're running the latest version
[netdisco2@netmanage ~]$ ~/bin/localenv cpanm --notest App::Netdisco
App::Netdisco is up to date. (2.033004)
And all the other devices we have seem to list their ports fine,
although we don't have any other kit with anywhere near as many
ports
as the 6509
Cheers
Dan
------------------------------------------------------------------------------
_______________________________________________
Netdisco mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
------------------------------------------------------------------------------
_______________________________________________
Netdisco mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
--- End Message ---
--- Begin Message ---
Hi Christian,
That's very nice! I´ve also tested it and it works fine! Thanks a mil for this
litte "christmas" present.
That makes the nexus discovery with vrf more easy and thus doesn´t require a
complex snmp config.
Sebastian
-----Ursprüngliche Nachricht-----
Von: Christian Ramseyer [mailto:[email protected]]
Gesendet: Mittwoch, 9. Dezember 2015 20:58
An: Sebastian Rösch <[email protected]>
Cc: Jeroen van Ingen <[email protected]>;
[email protected]
Betreff: Re: [Netdisco] Best way to discover Cisco Nexus N5K
On 08/12/15 13:42, Sebastian Rösch wrote:
> exactly, currently I have different snmp communites for each vrf.
> My favorite way would be to fetch the addresses with the ssh collector,
> because you there is a command in nexus for showing all vrf´s content:
> "show ip arp vrf all". That would be great... because I could remove all the
> additional devices I´ve just added for arpnipping the different vrf´s.
> But so far I wasn´t able to add another device type as described here, I´m
> not very familiar with all that perl stuff.....
> http://search.cpan.org/~oliver/App-Netdisco/bin/netdisco-sshcollector
>
> What do I need to do for this?
Hi Sebastian & List
Here is a module that I just made and tested on a N5kC5596Up. You must store
the file in lib/perl5/App/Netdisco/SSHCollector/Platform/NXOS.pm
(there's already all the other platform-specific .pm files there), then
reference it in deployment.yml like this:
sshcollector:
- hostname: 'n5.example.com'
user: whatever
password: topsecret
platform: NXOS
Maybe Olvier or another Netdisco dev can add it to the distro for future
releases.
Hope that helps
Christian
--- End Message ---
--- Begin Message ---
On 14/12/15 10:49, Oliver Gorwits wrote:
>
> ... I should add that there is no reason we could not improve the
> performance, but for a lack of time. I agree that it can and should be
> better!
>
> So if anyone would like to take a look at the DB queries, or submit
> patches for the web frontend, I would be very welcome indeed.
>
> To enable debugging to see the DB queries, run this in your terminal
> and visit Netdisco on port 5001:
>
> DBIC_TRACE=1 ~/bin/localenv plackup -p 5001 bin/netdisco-web-fg
>
Hi
I did in fact have a look at the issue a few days ago and tried the view
on some of our big boxes. What I have observed when viewing a 6500 with
some 300 ports and a bunch of columns like connected nodes enabled:
1. the actual SQL is not too bad, it finishes after maybe 10 seconds
2. the real problem seems to be the huge nested hash that gets assembled
by DBIx::Class when we call $set->all, it takes around 40 seconds to
build and makes the starman process grow to 1.3GB for my device. However
I was not able to get the search failed message in any case, maybe it's
just a question of having enough RAM/swap or liberal ulimits for the
netdisco user.
Unfortunately I have no practical experience with DBIx::Class. The
manual mentions HashRefInflator
(http://search.cpan.org/~ribasushi/DBIx-Class-0.082820/lib/DBIx/Class/ResultClass/HashRefInflator.pm)
as a lightweight way to access large results, maybe just switching to
this would be good enough.
Another harder to implement, but probably more promisiong solution would
be to only fetch the port names first, sort it, and then only retrieve
the full DBIx object trees for the interfaces on the current page.
Let's see, maybe I can have a look over these annoyingly long holidays
that are coming up :) Otherwise I hope this information might be helpful
to someone.
Christian
--- End Message ---
--- Begin Message ---
Hello Oliver,
Could you please inform me any update regarding to this issue , have you
already analyzed the information we provided you ?
Many thanks！
Best regards,
陈 杰 Jacy
GTAC Datacom Engineer
EU Support Center
HUAWEI European Enterprise Business Group
E-mail: [email protected]<mailto:[email protected]>
Tel: +80 03 38 88 888 (for West Europe)
+80 07 77 78 899 (for Poland/Norway/Sweden)
+40 31 22 61 900 / +40 31 22 61 907 / +40 31 22 61 909 /
+40 31 22 91 238 (for all EU countries)
Address: The Lake View 301-311 Barbu Vacarescu Boulevard,
2nd district, 020276 Bucharest Romania
enterprise.huawei.com<https://email-uk.huawei.com/owa/UrlBlockedError.aspx>
[cid:[email protected]]<https://email-uk.huawei.com/owa/UrlBlockedError.aspx>
[cid:[email protected]] <http://www.weibo.com/hwenterprise>
[cid:[email protected]]
<http://www.linkedin.com/groups/Huawei-Enterprise-4070523>
[cid:[email protected]]
<https://email-uk.huawei.com/owa/UrlBlockedError.aspx>
[cid:[email protected]]
This e-mail and its attachments contain confidential information from HUAWEI,
which is intended only for the person or entity whose address is listed above.
Any use of the information contained herein in any way (including, but not
limited to, total or partial disclosure, reproduction, or dissemination) by
persons other than the intended recipient(s) is prohibited. If you receive this
e-mail in error, please notify the sender by phone or email immediately and
delete it!
-----Original Message-----
From: Zhangyuanyuan (D)
Sent: 2015年11月24日 14:59
To: Oliver Gorwits; Chenjie (Jacy, Enterprise Romania GTAC);
[email protected]
Cc: Zoujing (Jingo Zou, DCLMT); Wuwei (Bryan); Zhaoshaoqi
Subject: 答复: [Netdisco] The connected device information is incorrect on
Netdisco.
Hi oliver,
We have two device: "192.168.1.247" and "192.168.1.246". Here is the output of
the related commands. It seems that all informations are correct.
--------------------------------------------------------------------------------------------------------------------------------------------------
192.168.1.247
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.247 -e c_id
[6870] 2015-11-24 17:43:12 info show: started at Tue Nov 24 12:43:12 2015 \ {
169046155.80.1 "e4:68:a3:f9:17:81",
169046155.92.1 "e4:68:a3:f9:17:81",
169046155.115.1 "00:6d:88:34:87:01"
}
[6870] 2015-11-24 17:43:12 info show: finished at Tue Nov 24 12:43:12 2015
[6870] 2015-11-24 17:43:12 info show: status done: Showed c_id response from
192.168.1.247.
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.247 -e c_ip
[6876] 2015-11-24 17:43:32 info show: started at Tue Nov 24 12:43:32 2015 \ {
169046155.80.1 "3.3.3.3",
169046155.92.1 "3.3.3.3",
169046155.115.1 "2.2.2.2"
}
[6876] 2015-11-24 17:43:33 info show: finished at Tue Nov 24 12:43:33 2015
[6876] 2015-11-24 17:43:33 info show: status done: Showed c_ip response from
192.168.1.247.
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.247 -e c_if
[6895] 2015-11-24 17:44:17 info show: started at Tue Nov 24 12:44:17 2015 \ {
169046155.80.1 80,
169046155.92.1 92,
169046155.115.1 115
}
[6895] 2015-11-24 17:44:18 info show: finished at Tue Nov 24 12:44:18 2015
[6895] 2015-11-24 17:44:18 info show: status done: Showed c_if response from
192.168.1.247.
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.247 -e c_port
[6897] 2015-11-24 17:44:24 info show: started at Tue Nov 24 12:44:24 2015 \ {
169046155.80.1 "GE2/0/31",
169046155.92.1 "GE2/0/19",
169046155.115.1 "40GE1/0/2"
}
[6897] 2015-11-24 17:44:25 info show: finished at Tue Nov 24 12:44:25 2015
[6897] 2015-11-24 17:44:25 info show: status done: Showed c_port response from
192.168.1.247.
--------------------------------------------------------------------------------------------------------------------------------------------------
192.168.1.246
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.246 -e c_id
[6872] 2015-11-24 17:43:19 info show: started at Tue Nov 24 12:43:19 2015 \ {
241393620.172.1 "e4:68:a3:f9:17:81"
}
[6872] 2015-11-24 17:43:20 info show: finished at Tue Nov 24 12:43:20 2015
[6872] 2015-11-24 17:43:20 info show: status done: Showed c_id response from
192.168.1.246.
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.246 -e c_ip
[6874] 2015-11-24 17:43:25 info show: started at Tue Nov 24 12:43:25 2015 \ {
241393620.172.1 "3.3.3.3"
}
[6874] 2015-11-24 17:43:26 info show: finished at Tue Nov 24 12:43:26 2015
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.246 -e c_if
[6893] 2015-11-24 17:44:11 info show: started at Tue Nov 24 12:44:11 2015 \ {
241393620.172.1 172
}
[6893] 2015-11-24 17:44:12 info show: finished at Tue Nov 24 12:44:12 2015
[6893] 2015-11-24 17:44:12 info show: status done: Showed c_if response from
192.168.1.246.
netdisco@dc-lmt:~$ ~netdisco/bin/netdisco-do show -d 192.168.1.246 -e c_port
[6899] 2015-11-24 17:44:32 info show: started at Tue Nov 24 12:44:32 2015 \ {
241393620.172.1 "40GE2/0/2"
}
[6899] 2015-11-24 17:44:33 info show: finished at Tue Nov 24 12:44:33 2015
[6899] 2015-11-24 17:44:33 info show: status done: Showed c_port response from
192.168.1.246.
netdisco@dc-lmt:~$
Best regards.
-----邮件原件-----
发件人: Oliver Gorwits [mailto:[email protected]]
发送时间: 2015年11月23日 17:10
收件人: Chenjie (Jacy, Enterprise Romania GTAC);
[email protected]<mailto:[email protected]>
抄送: Zoujing (Jingo Zou, DCLMT); Zhangyuanyuan (D); Wuwei (Bryan); Zhaoshaoqi
主题: Re: [Netdisco] The connected device information is incorrect on Netdisco.
Hi Jacy,
On 2015-11-17 08:44, Chenjie (Jacy, Enterprise Romania GTAC) wrote:
> Please be informed that we tried to contact support from Netdisco but
> nobody reply us, I am wondering if you have any other ways to contact
> them since we are not familiar with this Netdisco company.
I already replied to you with the following request:
Thank you for reporting this issue. To help us work out where the cause may be,
please could you provide the output of the following commands:
~netdisco/bin/netdisco-do show -d huawei_1 -e c_ip ~netdisco/bin/netdisco-do
show -d huawei_1 -e c_if ~netdisco/bin/netdisco-do show -d huawei_1 -e c_port
Many thanks,
regards,
oliver.
>
> Many thanks！
>
> Best regards,
>
> Jacy,
>
> FROM: Zhangyuanyuan (D)
> SENT: Wednesday, November 11, 2015 9:46 AM
> TO:
> [email protected]<mailto:[email protected]>
> CC: Chenjie (Jacy, Enterprise Romania GTAC); Zhaoshaoqi; Zoujing
> (Jingo Zou, DCLMT); Wuwei (Bryan)
> SUBJECT: The connected device information is incorrect on Netdisco.
>
> Hi, all
>
> I met a problem with Netdisco, I had sent a email last week. But there
> was no response.
>
> Could you help us analyze this issue. I will be deeply appreciated if
> this issue can be solved.
>
> The problem is that:
>
> when the network is consisted of Huawei Cloudengine Switches, the
> connected device information is incorrect on Netdisco, It’s
> inconsistent with the Topology.
>
> We have checked the packets that Huawei Switch sends to Netdisco, The
> SNMP packets were correct, the LLDP information in SNMP packet was
> consistent with the Topology. But, Netdisco showed wrong port
> connection.
>
> Here is the detail for this issue:
>
> The LLDP information on Huawei Cloudengine is as follows:
>
> <huawei_1>display lldp neighbor brief
>
> Local Interface Exptime(s) Neighbor Interface Neighbor Device
>
>
> ----------------------------------------------------------------------
> ---------------------
>
>
> 40GE2/0/2 103 40GE1/0/2 huawei_2
>
> GE2/0/19 114 GE2/0/31 huawei_1
>
> GE2/0/31 114 GE2/0/19 huawei_1
>
> We can see that, GE2/0/19 is connected to GE2/0/31. GE2/0/31 is
> another port on this device.
>
> But, on Netdisco, we can see that GE2/0/19 is connected to itself.
>
> 1. The implementation of LLDP on Huawei device is according to the
> protocol “802.1AB-2005”, The LLDP MIB description is as follows:
>
> 2. Huawei device use MIB node
> “lldpRemoteSystemsData(1.0.8802.1.1.2.1.4)” for LLDP, In this node
> “lldpRemPortId(1.0.8802.1.1.2.1.4.1.1.7)” is port’s neighbor
> information；
>
> 3. According to protocol, “lldpRemPortId” is the remote port that
> this device connects to.
>
> “lldpRemPortId” is string type:
>
> 4. In the SNMP packet that Huawei Switch sents, the value of the
> “lldpRemPortId” is correct.
>
> Take port GE2/0/19 for example, The value of “lldpRemPortId” is as
> bellow:
>
> lldpRemPortId.169046155.80.1 = 47.45.32.2f.30.2f.33.31
> //”169046155.80.1“ represents GE2/0/19 ” ,and
> “47.45.32.2f.30.2f.33.31” represents G.E.2./.0./.31
>
> lldpRemPortId.169046155.92.1 = 47.45.32.2f.30.2f.31.39 //”
> 169046155.92.1” represents GE2/0/19 “, and “47.45.32.2f.30.2f.31.39”
> represents G.E.2./.0./.19.
>
> lldpRemPortId.169046155.115.1 = 34.30.47.45.31.2f.30.2f.32
--- End Message ---
------------------------------------------------------------------------------
_______________________________________________
Netdisco mailing list - Digest Mode
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
