What could cause RVRD to use a lot of memory and appear like it has a memory leak?
Article ID: KB0085551
Updated On:
| Products | Versions |
|---|---|
| TIBCO Rendezvous | - |
| Not Applicable | - |
Description
Resolution:
It is entirely normal for the amount of memory being used by RVRD to grow (possibly significantly) for the following reasons:
1. The (physical) connection between two neighbors has limited bandwidth, causing the sending RVRD to buffer data pending delivery to its neighbour.
2. If there are attached publishers, the RVRD will have to grow in order to maintain the 60-second reliability interval (remember: RVRD can also act as an RVD)
3. If there are attached subscribers who are slow consumers, the RVRD will grow as it buffers data pending delivery to its attached subscribers.
4. A dirty network (i.e. one in which retransmission rates are high) can also contribute to RVRD (or RVD) growth because the daemon will hold on to missed packets pending receipt of re-requested data.
5. RVRD will also allocate roughly 10MB for each router instance, which has been defined. If there are a large number of router instances defined, the basic size of RVRD (without the other issues described above) can be quite large.
The issue may be something as simple as not having sufficient memory allocated to the user who is running RVRD. Or, if the connection between neighbours has limited bandwidth, enabling the maximum backlog feature of RVRD may mitigate the situation.
It is entirely normal for the amount of memory being used by RVRD to grow (possibly significantly) for the following reasons:
1. The (physical) connection between two neighbors has limited bandwidth, causing the sending RVRD to buffer data pending delivery to its neighbour.
2. If there are attached publishers, the RVRD will have to grow in order to maintain the 60-second reliability interval (remember: RVRD can also act as an RVD)
3. If there are attached subscribers who are slow consumers, the RVRD will grow as it buffers data pending delivery to its attached subscribers.
4. A dirty network (i.e. one in which retransmission rates are high) can also contribute to RVRD (or RVD) growth because the daemon will hold on to missed packets pending receipt of re-requested data.
5. RVRD will also allocate roughly 10MB for each router instance, which has been defined. If there are a large number of router instances defined, the basic size of RVRD (without the other issues described above) can be quite large.
The issue may be something as simple as not having sufficient memory allocated to the user who is running RVRD. Or, if the connection between neighbours has limited bandwidth, enabling the maximum backlog feature of RVRD may mitigate the situation.
Issue/Introduction
What could cause RVRD to use a lot of memory and appear like it has a memory leak?
Feedback
Yes
No
Powered by
