MDS R81.20 on KVM, fresh install broken.

[LukeDRussell1]

Hello,

I'm unable to get an MDS install working using this qcow2 image. I've been using this image for Gateway and SMS for a couple of months in my lab, everything seems to work fine. With MDS, the first symptom I noticed is I can't connect with Smart Console. SSH and Web to Gaia work fine though.

When I tried to run `api status` I would get an error about the missing file /opt/CPsuite-R81.20/fw1/conf/cpmServerSettings.props. I copied this from an SMS server to get past that error.

I also tried updating with the latest Take, but the Gaia web gui errors and suggests running `cpstop` but that doesn't work on an MDS server. 
I've tried about 5 times with a fresh VM, now I'm tearing my hair out.

Some probably relevant outputs that might help:

[Expert@cp-mds:0]# mdsstat

CPM: Check Point Security Management Server is not running

+------+--------------------+-----------------+-------------+-------------+-------------+-------------+-------------+
| Type | Name               | IP address      | FWM         | FWMHA       | FWD         | CPD         | CPCA        |
+------+--------------------+-----------------+-------------+-------------+-------------+-------------+-------------+
| MDS  |          -         | 172.30.32.10    | down        | down        | down        | down        | down        |
+------+--------------------+-----------------+-------------+-------------+-------------+-------------+-------------+
Total Domain Management Servers checked: 0     0 up   0 down
Tip: Run mdsstat -h for legend

[Expert@cp-mds:0]# mdsstart 
Starting cpWatchDog
Starting CPM Server ...
[1] 8715
CPM Server is running.
Start Search Infrastructure...
index mode was set to true
startsearch: dbsync does not run on Multi-Domain Security Management
cpwd_admin: 
Process SOLR started successfully (pid=9230) 
Starting RFL ...
cpwd_admin: 
Process RFL started successfully (pid=9262) 
Starting SmartView ...
Starting SmartView...
cpwd_admin: 
Process SMARTVIEW started successfully (pid=9311) 
Start Log Indexer...
cpwd_admin: 
Process INDEXER started successfully (pid=9594) 
Start SmartLog Server... 
cpwd_admin: 
Process SMARTLOG_SERVER started successfully (pid=9806) 

No need to run Adjuster Service - no clients were found
Starting Log Indexer...
[1]  + Done                          /opt/CPsuite-R81.20/fw1/scripts/ngm_start.sh
/opt/CPmds-R81.20/customers: No such file or directory.

[Expert@cp-mds:0]# api status

API Settings:
---------------------
Accessibility:                      Unknown
Automatic Start:                    Unknown

Processes:

Name      State     PID       More Information
-------------------------------------------------
API       Stopped   0         
CPM       Stopped   0         
FWM       Stopped   0         
APACHE    Started   8507      

Port Details:
-------------------
JETTY Internal Port:               0
JETTY Documentation Internal Port: 0
APACHE Gaia Port:                  443

Profile:
-------------------
Machine profile:                   Medium env resources profile
CPM heap size:                     1280m

                          Apache port retrieved from: httpd-ssl.conf


--------------------------------------------
Overall API Status: The API Server Is Not Running!
--------------------------------------------

API readiness test FAILED. The server is down and unable to receive connections!

Notes:
------------
To collect troubleshooting data, please run 'api status -s <comment>'

[PhoneBoy]

I suspect these qcow images are not set up to run MDS, which has different requirements than a regular SMS (more disk/RAM, NIC configuration).
What are the specs on the VM you’re attempting to deploy this on? (RAM/CPUs/Disk/NICs)

[LukeDRussell1]

It will be easy enough for me to give it more resources. I can't remember where I found the requirements, but I'm currently running 4vCPU, 6 GB, virtio NICs. I'm not sure how big the disks were. One attempt I set the data disk to 100 GB manually.

I'd be happy to take a suggestion on sizing. 

[the_rock]

To me, that does not look like the right image for eve-ng. Did you follow below link to make sure naming is right? I know for mds, you would use same image as you were installing regular mgmt. I tested that before in eve-ng, no issues.

Andy

https://www.eve-ng.net/index.php/documentation/qemu-image-namings/

[LukeDRussell1]

I'm not using eve-ng.

[the_rock]

Sorry my bad for assuming so. Glad you got it working.

Andy

[LukeDRussell1]

I'm using Cisco Modelling Labs, which is built on top of KVM.

I also contributed some configs to CML-Community repo in case anyone else wants to run it. I'll add notes in that about increasing the resources for MDS. 

[PhoneBoy · Accepted solution]

That's barely enough to run a gateway.
MDS needs a lot more resources (at least 32GB RAM, 8 Cores).
Refer to the requirements here: https://sc1.checkpoint.com/documents/R81.20/WebAdminGuides/EN/CP_R81.20_RN/Content/Topics-RN/Open-Se... 

[LukeDRussell1 · Accepted solution]

I increased the specs in line with the Open Server minimums for MDS (8c, 32GB, 150 GB disk) and it works immediately!

Thanks @PhoneBoy