Dear All,
Not sure how many of you out there are using Opsware but this is the first time I am seeing this product.
Of late the place where I am working is evaluating HP Opsware product for file distribution etc. This piece of sw seems to work like master & agent. During the evaluation we have found that once the agent is installed on the AIX server, people who are logged on to opsware master console, can gain access to the system as root and do whatever they want.. The way the agent was installed was, by adding the AIX server in question as an object and pushing the agent to the AIX server. A non-root user having full sudo access has been given while pushing the agent, so the agent installation can happen. When we have asked the Team working from Opsware about this, they said they will have a workaround and limit the access to people who are going to work on the console. My questions are
1. So there is a superuser in Opsware console which can restrict access, but this superuser can logon to any AIX server without a password or any sort of authentication. Worst is the login from opsware console does not leave any audit trail in the OS logs. I have verified the ssh logs and wtmp to see if they report root user being connected, but could not see any info.
2. Since the agent is installed by Opsware which had the legitimate access, does it mean for all the rest of the time that piece of sw can logon without any authentication. Is this right???
3. If we escalate this to IBM, shouldn't IBM be obliged to fix this by not allowing third party applications to do this kind of stuff on their OS. Say if there was a vulnerability for sendmail, doesn't IBM come out with a fix or not...
Any light on this is very much appreciated.
Regards
KPKIND
Not sure how many of you out there are using Opsware but this is the first time I am seeing this product.
Of late the place where I am working is evaluating HP Opsware product for file distribution etc. This piece of sw seems to work like master & agent. During the evaluation we have found that once the agent is installed on the AIX server, people who are logged on to opsware master console, can gain access to the system as root and do whatever they want.. The way the agent was installed was, by adding the AIX server in question as an object and pushing the agent to the AIX server. A non-root user having full sudo access has been given while pushing the agent, so the agent installation can happen. When we have asked the Team working from Opsware about this, they said they will have a workaround and limit the access to people who are going to work on the console. My questions are
1. So there is a superuser in Opsware console which can restrict access, but this superuser can logon to any AIX server without a password or any sort of authentication. Worst is the login from opsware console does not leave any audit trail in the OS logs. I have verified the ssh logs and wtmp to see if they report root user being connected, but could not see any info.
2. Since the agent is installed by Opsware which had the legitimate access, does it mean for all the rest of the time that piece of sw can logon without any authentication. Is this right???
3. If we escalate this to IBM, shouldn't IBM be obliged to fix this by not allowing third party applications to do this kind of stuff on their OS. Say if there was a vulnerability for sendmail, doesn't IBM come out with a fix or not...
Any light on this is very much appreciated.
Regards
KPKIND