A client of our runs several workstations and a server with PDM.
The thing is, the client wants the check in and out procedure to be faster than the time it takes now (couple of minutes if I understood correctly). The local support for Solidworks has already said that was all the performance they could get, but the client is stubborn. :-/ Don't ask..
Where could be a possible bottleneck in all this?
Network is 1Gb/s, confirmed through file copy, workstations run 2 SSD's in RAID0, and the server has two 15K disks RAID1 for the database, RAID5 for the archive.
The server is basically doing nothing all day, no sign of any limit reached, same for clients.
Hardware upgrades to 10G network and full SSD storage in the server are no problem if it really makes a difference. (they're currently running on a temporary server which is hooked up to 100M network and they haven't complained about any slowdown)
Is there perhaps a list showing each step of what happens during check in and out? (disk activity, calculations, network activity, database activity, etc.)
Any ideas are greatly appreciated.
This situation, and unfortunately not for the first time is familiar to me.
Moreover - now I have the same problem.
Once at me it turned out to correct something, but as definitely I can't tell.
I then much everything tried:
- first of all - cleared remote files (on a tab of remote files)
- checked errors of export of the vaults settings (errors shan't be)
- checked the SQL lists for cards
- cards, templates and cards of templates
- settings of notifications in workflows
- settings on the SQL Server
- settings of all users directly and everyone, groups (menu, adding of files and setup of viewing)
- settings of a fayevol and antivirus (ports, exceptions, rules)
- deleted a garbage of test storages in the register on the server
- the rights of the user as administrator for everyone the *.ekh (SW,PDM...)
and many other things
At the moment the problem is again - but there is no decision yet.
But there is some information which can be useful to someone and will help to solve something.
1. At me it is a problem of specific vaults (perhaps the reason recent changes of settings)
2. The problem is watched (at least at me) on specific computers and users.
If I come under other user - problems can not be.
That the most interesting - in case of an input from other computer - the same user has no problem.
Cleaning of a local cache gives nothing.
3. If to create the new user - having copied settings old - problems isn't present too (but it not the decision).
4. There are suspicions that recent failures in operation of the server can be the cause (some settings
the user are spoiled on the server?) - but what with it to do not clearly.
5. Perhaps the reason - incorrect or conflictual new conditions for categories and workflows.
Changing different conditions - it turned out to change delay period several times (and directly on several computers and at different users at the same time)
6. It is interesting that problems are watched irrespective of the sizes of files, their locations, but files with some extensions can not have problems.
7. Problems are watched as a rule in case of large volumes of files in storage and at more active users.
- the diagram of cold storage wasn't set up
- the conductor of objects isn't used
- the task list - there are no errors
And still - several users were deleted recently (it was necessary to clean mailing lists of notifications) , possibly cleaned not everything (there could be some records and communications on the SQL server - and manual cleaning is necessary).
I ask to excuse for machine translation.