Imported from previous forum
I have a question on the session logs maintained by Quickfixj. We have FIX engine built on webMethods Integration servers and there are two instances of FIX engine running on two servers in parallel. The session log files are shared between the FIX engines so that the failover can be handled seamlessly. The FIX engines are configured to start up at 7:00 AM and they shut down at 7:00 PM. The session logs are not deleted at the end of the day as Quickfixj can handle the session startup and shutdown on its own. We observed that sometimes when the session comes up in the morning, the sequence numbers starts at 1 but does not increment. That is, the FIX engine sends sequence no 1 to the client followed by sequence no reset message continuously, until the session log files are manually deleted and the FIX session is restarted. This happens at random, sometimes after being active for 3 days and sometimes after 15 days.
The logs look like,
2011-04-08 07:30:20,955|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:20.95556=*****10=062
2011-04-08 07:30:31,113|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:31.11356=*****10=050
2011-04-08 07:30:42,628|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:42.62856=*****10=063
2011-04-11 07:30:00,887|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:00.88756=*****10=058
2011-04-11 07:30:21,046|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:21.04656=*****10=048
2011-04-11 07:30:31,201|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:31.20156=*****10=042
2011-04-11 07:30:41,544|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:41.54456=*****10=053
2011-04-12 07:30:00,875|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:00.87556=*****10=056
2011-04-12 07:30:21,030|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:21.02956=*****10=050
2011-04-12 07:30:31,182|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:31.18256=*****10=051
2011-04-12 07:30:41,337|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:41.33756=*****10=054
Any suggestions or comments will be helpful. Is this a known issue? Should we be using individual session log files for each FIX engine? Or is there anything else that we can code for to avoid this? Thanks in advance.
I have a question on the session logs maintained by Quickfixj. We have FIX engine built on webMethods Integration servers and there are two instances of FIX engine running on two servers in parallel. The session log files are shared between the FIX engines so that the failover can be handled seamlessly. The FIX engines are configured to start up at 7:00 AM and they shut down at 7:00 PM. The session logs are not deleted at the end of the day as Quickfixj can handle the session startup and shutdown on its own. We observed that sometimes when the session comes up in the morning, the sequence numbers starts at 1 but does not increment. That is, the FIX engine sends sequence no 1 to the client followed by sequence no reset message continuously, until the session log files are manually deleted and the FIX session is restarted. This happens at random, sometimes after being active for 3 days and sometimes after 15 days.
The logs look like,
2011-04-08 07:30:20,955|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:20.95556=*****10=062
2011-04-08 07:30:31,113|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:31.11356=*****10=050
2011-04-08 07:30:42,628|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110408-11:30:42.62856=*****10=063
2011-04-11 07:30:00,887|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:00.88756=*****10=058
2011-04-11 07:30:21,046|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:21.04656=*****10=048
2011-04-11 07:30:31,201|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:31.20156=*****10=042
2011-04-11 07:30:41,544|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110411-11:30:41.54456=*****10=053
2011-04-12 07:30:00,875|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:00.87556=*****10=056
2011-04-12 07:30:21,030|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:21.02956=*****10=050
2011-04-12 07:30:31,182|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:31.18256=*****10=051
2011-04-12 07:30:41,337|INFO|Admin Message sent to counter party: 8=FIX.4.49=5235=534=149=*****52=20110412-11:30:41.33756=*****10=054Any suggestions or comments will be helpful. Is this a known issue? Should we be using individual session log files for each FIX engine? Or is there anything else that we can code for to avoid this? Thanks in advance.
To update, We turned off the FIX engine on one of the instance and let it run on only one server. It ran fine for the past 20 days and we had the sequence number issue happen again today. That is the sequence number started at 1:00 today morning but it did not increment after that. We had to delete the session logs and restart the FIX engine to fix this. Is this a limitation in quickfixj framework? Is there a workaround that we can try?
Thanks!!