Welcome to the Cisco Networking Professionals Ask the Expert conversation. This is an opportunity to discuss with Cisco experts Bruce Merchant and Jossy Antony about Cisco Works LMS implementation. Bruce Merchant is a technical marketing engineer supporting CiscoWorks LMS. During the eight years he has been with Cisco Systems, Inc. he has worked with routing and switching, wireless, IP telephony, and mobile IP. Mr. Merchant provides training, demonstrates different technologies, and provides assistance in customer implementations. Jossy Antony is a technical marketing engineer at Cisco Systems Inc. He has been with Cisco for over five years and is currently part of the Network Management Technology Group. The products he supports include CiscoWorks LAN Management Solution and CiscoWorks Small Network Management Solution.
Remember to use the rating system to let Bruce and Jossy know if you have received an adequate response.
Bruce and Jossy might not be able to answer each question due to the volume expected during this event. Our moderators will post many of the unanswered questions in other discussion forums shortly after the event. This event lasts through March 10, 2006. Visit this forum often to view responses to your questions and the questions of other community members.
how many diffrent servers are required to run LMS correctley, and how should he modules be split between the servers?
LMS can run on a single server, either Windows or Solaris. It is also possible to have multiple servers, with one as master database and the others as subsidiary.
Whether you need multiple servers is determined by the number of devices (routers, switches, etc.) to be managed. One server with all LMS applications can manage 1500 - 2000 devices. If more than that need to be managed, you can either split the network by geography, network hierarchy, organizational division, however you want; or you can distribute different applications across separate servers, with each tied back to the master database. A server running a single application (such as RME or Campus Manager) can manage up to 5000 devices. Note one exception - the limit for a server running DFM is 1500-2000 devices.
The recommendations for server sizing are contained in the Quick Start Guide on CCO at
The guidelines for implementing multiple servers are available as a white paper - the LMS Large Scale Deployment Guide. This is availble on an internal website; please contact your local Cisco contact to obtain it.
The minimum Wintel server requirement for LMS 2.5.1 is
CPU: > 1GHz
I have a Wintel server with 2 x 700 MHz CPU on board and 3 GB RAM. Can this server be used to install LMS 2.5.1?
Hi, I need to Cisco recommendation of deploying the LMS and VMS on same Unix server. the network contains 200 nodes for LMS and 9 FWSM/IDSM
Latest version of LMS (2.5) and VMS 2.3 cant co-exist on same server; you have to use two separate servers for both applications.
Old version of LMS (2.2) can co-exist with VMS 2.3 and you looking for a solution around that, you could use a SUN V 240 server with 2GB RAM.
CiscoView does not display all the cards from all (about 10) of our Cisco-7304 devices.
We have installed CiscoView 6.1 with all devices packages from CCO.
I have reinstalled the router7000 package, but it continues doing the same.
- 7304 IOS: 12.2(25)S3.
- Operating System: Solaris 9 (5.9).
- CiscoView version: 6.1.
All our cisco-7304 have a card with "CardType=1568" (Network Service Engine 100 Daughter Card") that is not supported in the "Readme file" of the Device Package of the cisco-7000 routers downloaded from CCO, but I think this card may not affect the display of the network cards.
CiscoView only display a card with 3 ethernet ports.
Please, Could you help me?
We need more information to troubleshoot your issue with CiscoView
Please provide following information
What version of the CiscoView device package for Rtr7000 was installed?
output to know which cards are installed on the system
+ Snmp walk for the following mibs:
+ Enable snmp trace and activity trace for CiscoView, open the chassisview for the device. Take the following logs:
Please, there is a way to send you all this information through a more safe way than this clear forum (an email ...)
Thanks a lot.
How about opening a Service request with Cisco TAC to troubleshoot your problem. Or give me your email ID I will contact you
You'll want to give a good chunk of disk to:
/var (syslog, software images and log files go here)
/opt (application and databases! go here
I'd also recommend a separate
I used to recommend simply a separate directory in /opt for the LMS backups (ie. /opt/CWbackup), but I noticed that a few customers got BIIIIG backups and when those backups ate all the space in /opt, then LMS would stop working (no more space for the databases to grow). If you have a separate partition for the backups, if they fill that partition, LMS will still run (without doing backups).
Parition sizes depend on:
1) how many devices you are managing
2) what other applications might go on this server
3) how many backups you'll retain (I've seen a customer with 2GB backups per day - how many days will you store?)
4) how many configs and software images you'll store
Considering that you have two 80-GB drives. If this is for CiscoWorks LMS only, consider mirroring the drives and doing this break-down:
swap 8GB (yes, this is not a typo - see the install guide)
These drive will be mirrored and I am currently monitoring 500 devices, 1 backup then to tape. Thank you for the recommendation on the slices.
Here the recommendation for a single 80 GB disk
/ 8192 MB (8 GB - anything not allocated in the other partitions comes from /, so make it a little bigger than recommended)
/var 16384 MB (16 GB - Syslog, device images, config files... etc gets stored here)
/opt 32768 MB (32 GB - application goes here /opt/CSCOpx)
swap 8192 MB (8 GB - conforms to our 8 GB swap rule)
/usr 10240 MB (10 GB - in case we get some stuff in /usr/local)
/export/home 6144 MB
And you have dual disk move /var to second disk
I'm reinsatlling LMS and need to know
what OS it will perform best Windows or Solaris
hardware is Dell2850 Xeon 2.8Ghz dual
4*73GB scsi in RAID5
LMS is not supported on Solaris running on x86 platforms, Solaris OS supports for LMS is limited to SUN SPARC (IIIi and IIIcu processor based systems)
I would recommend you to use Windows 2003 server with SP1 on your Dell Platform
Can I have pseudo high availability, by installing redundant LMS 2.5 components on two different servers, e.g. Server 1: CS master, RME, CM, CiscoView, Server 2: CS slave, RME, CM, CiscoView? I don't have problem with double polling. I just want DCR to sync. Is DCR sync'ing one-way (master2slave) or bi-directional? Would licensing be for two servers in this setup? If master goes down, and I add/del devices in the slave DCR, would the changes be sync'ed from slave when master comes back up?
The short answer is yes - you can have 2 servers with duplicate applications, running in parallel, doing duplicate data gathering and management. You would have to purchase 2 LMS licenses. DCR syncing is one-way, from master to slave.
If master goes down, you can still access the applications running on the slave, and manage the network.
If the master will be down for a long time, you can promote the slave to stand-alone. Then you can make changes to the DCR, add new devices, etc. When the master comes back up,
You will have to change the slave back to slave, and associate it again with the master. However, credential changes or device deletions are not populated to the master when it comes back up but new devices will get added to master server. One way to update the credentials on master is take the export of devices and credentials from the slave when it is in standalone mode and import into master DCR
If the master will be down for a long time, you can promote the slave to new Master. Then you can make changes to the DCR, add new devices, etc. When the master comes back up, change the old master to slave and associate with new master. In this method changes you made will get reflected on both servers.
I forgot about IPM. Can it be installed in this redundant fashion as well? Off the top of my head, I thought I read somewhere running more than one instance of IPM 2.5.x with the same set of operations (sources/targets) could cause problems. Is this still the case with IPM 2.6, and does this preclude running parallel IPM instaces?
IPM also support hot and standby mode (redundant), but there going to be two sets of collectors on the routers doing the same operations.
LMS 2.5.1 doesnt support default credentials or Group based credentials. When new devices added to DCR by campus discovery or 3rd party NMS import all credentials except SNMP read community will be null. Next version of LMS we are trying to address this issue by having default credentials which can be override by each device level credentials.
I just purchased Cisco Works LMS 2.5.1 and I found out that the Topology Services on the Campus Manager doesnt display the network events in real time, so if a router goes out it doesnt display the icon in red color.
Is there something that I can configure so I can get this functionality??
I appreciate your comments!
The main function of Campus manager Topology Services is to provide you dynamic network documentation.
The topology-services tool allows you to
View detailed network information by devices, links, and ports
Display physical and logical services
Segment your network logically and manage workgroups that use VLANs
View port, device, and trunk attributes along with port information by VTP domain for simplifying VLAN configuration
View reports on inconsistencies or misconfigurations of the physical or logical network setup
Campus Manager periodically polls devices and links to check the status and the default interval for polling is 2hrs. If you have a small network then you could change the Poll Interval from default 2hrs to 10 or 15 minutes to get more frequent status changes on the Topology map (this could take more CPU cycles on your server). Again I dont recommend you to use Campus manager as an availability monitoring tool. Device Fault Manager, another application in LMS does the availability monitoring
Here the steps to change Pooling Interval in Campus manager
1. Select Campus Manager > Administration.
2. Select Admin > Campus Data Collection > Schedule Data Collection.
3. Enable polling by selecting the check box and Change the value
4. Click Apply
I have installed LMS 2.5 with May 2005 updates package, and CM SP1 on it. Now at User Tracking End User report I can see the MAC addresses but not ip address and hostnames.You can see the situation in the attachment file.
How can i solve this problem?
Hi there i haev LMS 2.5.installed in our Network, but in the User trackibg Dicovery i am able to see only few Host on the Network,i have arougn 1500 hosts but i can able to see only 300 .can u please tell me how is these User trackign working and when it will display all the Hosts in the Network.
Any help great Appreciated,
Falkaabi, this may be related to your previous post where you said you are getting MAC addresses, but no IPs. If the 1200 missing devices are your routers, then there is no way for User Tracking to associate the MAC addresses back to IPs.
The way the process works is:
1) User Tracking (UT) obtains the device list from Device Credential Repository (DCR)
Caveat - if DCR doesn't contain reachable devices with usable SNMP community strings, we're in trouble
2) UT scans the SNMP MIB tables on the switches for the CAM/MAC address tables.
3) UT scans the SNMP MIB tables on the routers for ARP/IP cache - this gives us MAC to ARP/IP pairing.
4) UT makes requests to the DNS server with the IP address to reverse-resolve the hostname.
That's the process in a nut-shell.
So, you need to ensure you have as many routers in your inventory that are known and reachable before you can worry about the tables being populated.
Here's a good doc on troubleshooting UT.
Now, to this post where you aren't getting all your devices. Campus Manager has a network discovery component called ANI that finds the devices and feeds DCR. This discovery is based on CDP (Cisco Discovery Protocol) which is a Layer-2, SNAP-based adjacency broadcasting feature. If you have CDP disabled on your devices or if you have non-Cisco gear between your Cisco gear then we won't get all the discovery done. You CAN manually import devices.
You should check at your router/switches snmp configuration.
If you put configurations like "snmp-server view xxx" this will restrict the information Ciscoworks get from your switches.
I faced this problem once, where i can get the mac address but not the hostname and ip address. When i checked on the switch configuration.. there's snmp-server view configuration. When i removed them, i'm able to view the ip address and hostname in User Tracking.
Hopes this helps.
Which is the switch model these devices connected to, if it is a 3550-24 upgrade the IOS to 12.1(22).
Try applying LMS 2.5 December Update which has fixed many bugs in LMS 2.5