Description:
If you notice rvd/rvrd run out of memory problem, and also notice some DATALOSS problem, you may run into this defect.

1-81EIPF: Fixed a daemon defect in which data loss could result in unreleased memory (in receiving daemons). In situations with high concurrency, this memory growth could rapidly consume all available process storage.

OS Version: ALL
RV version: Since RV 7.2 till RV 7.5.2 hotfix HF03, it is fixed in 7.5.2HF04 and 7.5.3.

Symptoms:

A. rvd run out of memory or rvd memory footprint is unexpected high. Most of time,
you may notice "rvd: DrainQueue: out of memory"/"rvrd: DrainQueue: out of memory" with your demons

B. DATALOSS advisory messages

To Identify this problem:

enable undocumented Profiling through the HTTP interface by simply going to:

http://&lthostname>:&ltport>/profiling

From that page, check "Memory" boxes and then "submit".  
The profiling data will then be outputted to the rvd/rvrd logfile. If you do not use log file with your daemon, please click
"Current Log" in left side of your daemon HTTP interface main page.

Profiling could cause logfile grow quite significantly in size, please disable profiling after you captured several memory profiling data which covers the problem period.

You should get memory profiling similar to the following:

Email Body   : 2007-03-02 17:33:28 rvd: Memory:
2007-03-02 17:33:28 rvd:  Msg pool:                  848000           2000           1999              0              0.
2007-03-02 17:33:28 rvd:  Buffer pool:              1512000          21000          13298.
2007-03-02 17:33:28 rvd:  Packet pool:           1953792000        1272000         103214.
2007-03-02 17:33:28 rvd:  Accumulator pool:       302010624            288            216.
2007-03-02 17:33:28 rvd:  Queue entry pool:        53196000        1209000           5249.
2007-03-02 17:33:28 rvd:  Net pool:                 1048576           8192            440              0              0.
2007-03-02 17:33:28 rvd:  IPQ pools:                8419904             32             32              0              0
2007-03-02 17:33:28 rvd:                                               256            256            132             12.
2007-03-02 17:33:28 rvd:  Multicast send:           2179630.
2007-03-02 17:33:28 rvd:  Multicast receive:              0.
2007-03-02 17:33:28 rvd:  Unicast send:                   0.
2007-03-02 17:33:28 rvd:  From net:                 1203752.
2007-03-02 17:33:28 rvd:  To clients:                     0
*********************************************************************
2007-03-02 17:33:28 rvd:  Buffer pool:              1512000          21000          13298.
2007-03-02 17:33:28 rvd:  Packet pool:           1953792000        1272000         103214.
2007-03-02 17:33:28 rvd:  Accumulator pool:       302010624            288            216.
2007-03-02 17:33:28 rvd:  Queue entry pool:        53196000        1209000           5249.
2007-03-02 17:33:28 rvd:  Net pool:                 1048576           8192            440              0              0.
2007-03-02 17:33:28 rvd:  IPQ pools:                8419904             32             32              0              0
2007-03-02 17:33:28 rvd:                                               256            256            132             12.
2007-03-02 17:33:28 rvd:  Multicast send:           2179630.
2007-03-02 17:33:28 rvd:  Multicast receive:              0.
2007-03-02 17:33:28 rvd:  Unicast send:                   0.
2007-03-02 17:33:28 rvd:  From net:                 1203752.
2007-03-02 17:33:28 rvd:  To clients:                     0
*********************************************************************


If the values for "Packet pool:" and "From net:" are high, that means you ran into this defect.

7.5.3 fixed a daemon defect (1-81EIPF)in which data loss could result in unreleased memory.