d98rolb Posted January 27, 2012 Report Share Posted January 27, 2012 I have checked in about 8500 files and sometimes the update of pending changes is VERY slow. I have a refresh here that takes 8 minutes the first time. How is this possible ? In this case there is 182 files changed. Largest files are about 10 MB in size. I am using a local SQL Server Developer edition on a virtual machine with VMWare Fusion on a MacBook pro. Link to comment Share on other sites More sharing options...
manu Posted January 27, 2012 Report Share Posted January 27, 2012 Hello d98rolb, 8 minutes is not normal at all..... Do you have the option "Compare file contents instead of timestamps when determining "changed" status" enabled? It will be faster if you don't have it enabled. Let's see if the performance is dying at IO calls. Repeat the process with all the Pending changes view options disabled but the "Show checked-out items" tell us the time. Then repeat the test with the "Show checked-out items" and "Show changed items" options enabled. Manu. Link to comment Share on other sites More sharing options...
d98rolb Posted January 27, 2012 Author Report Share Posted January 27, 2012 Strange now it is normal times again, maybe 3-4 seconds for refresh. Seems to be slow when I not used PlasticScm for a while and come back and refresh. And that option in preferences has always been cleared as default. Link to comment Share on other sites More sharing options...
manu Posted January 27, 2012 Report Share Posted January 27, 2012 Hi d98rolb, The "pending changes view" only connects one time with the server and the call should be very very fast. The rest of the time is only IO disk time. But, just in case, please send to us the "ChannelCall.txt" log file, you will find it on the PlasticSCM server directory. You can reach me at: mlucio at codicesoftware dot com. Manu. Link to comment Share on other sites More sharing options...
d98rolb Posted January 31, 2012 Author Report Share Posted January 31, 2012 The slowness seems to be gone now No idea why that happened before. Link to comment Share on other sites More sharing options...
manu Posted January 31, 2012 Report Share Posted January 31, 2012 Fine, we will be alert if it raises again.... FYI, we are developing a new cool command that will be useful to determine where are the installation bottlenecks! It's very interesting due to we will be able to see the network performance, IO performance, CPU and so on! Link to comment Share on other sites More sharing options...
d98rolb Posted January 31, 2012 Author Report Share Posted January 31, 2012 "But, just in case, please send to us the "ChannelCall.txt" log file, you will find it on the PlasticSCM server directory." Checked that file and it is empty. Probably some setting to enabled logging ? Link to comment Share on other sites More sharing options...
manu Posted January 31, 2012 Report Share Posted January 31, 2012 Yes, can you send to me your "plastic.server.log" and your "loader.log.conf" files? Link to comment Share on other sites More sharing options...
d98rolb Posted January 31, 2012 Author Report Share Posted January 31, 2012 They are sent. I believe the slowness come from when I check in some files by mistake, in this case some log-files. I delete the files from disk afterwards. Then when I refresh pending changes it was very slow. The files was marked as missing from disk. So I marked the files in items and choose Update forced. Then the files was marked as Controlled/Cloaked instead. And the refresh was faster But I think PlasticScm should take care of this and not try to search for nonexisting files. Link to comment Share on other sites More sharing options...
manu Posted February 1, 2012 Report Share Posted February 1, 2012 Mmmmmmm That scenario should be faster too. If you can reproduce the long time issue I can connect with you to diagnose the problem. Link to comment Share on other sites More sharing options...
psantosl Posted February 1, 2012 Report Share Posted February 1, 2012 We're about to make public a new "cm iostat" command measuring disk io speed and network speed. Most of the time disk io is the culprit of all the slowdowns... Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.