Neverending Story (decadence1) wrote in lj_userdoc,
Neverending Story

FAQ #162 - Why can't I stay logged in?

Inspired by a recent lj_dev discussion this entry discusses FAQ #162 - Why can't I stay logged into LiveJournal, and what can I do to solve this problem?
The nature of this FAQ is that it discusses technical issues - cookies, session expiration etc. For several areas, choosing what's going to ultimately be the best text will depend on how those things do actually work. I'm not really confident I understand all of this (!) so input from others is appreciated.

If you are having problems staying logged into LiveJournal -- for instance, if you are randomly logged out when moving from page to page -- it is most likely a problem with your browser or your Internet connection. There are several steps you can take that may help to solve this problem.
Suggest: Remove 'help to'. Rationale: I think it's a little tautological. Also suggest striking 'your'.

The cache suggestion exists for the scenario where a user logs in and their cookie is set but they then attempt to access a page and it happens to be recently visited (before they logged in) and cached. As hitting Reload may help, I'd sooner just include that info in this FAQ; pointing to FAQ #160 for only if the problem persists and they want to clear their entire cache. Otherwise we're asking them to view another page full of info where they mightn't need to. The first paragraph of FAQ #160 explains the issue already talked about in this FAQ and the final paragraph is an unrelated issue. Consequently it's probably worth just giving the limited info they will need, in FAQ #162.
Suggest: Try hitting Refresh or Reload in your browser. If this does not help you may need to clear your browsers cache; instructions can be found in this HowTo journal article:, provided as a courtesy by support volunteers.

... Then, log out of LiveJournal, delete all LiveJournal cookies in your browser, and make sure that your browser is properly accepting cookies. Information on deleting cookies can be found at Information on making sure that your browser is accepting cookies can be found in your browser's help files.

That paragraph discusses manual cookie deletion (or at least specifies LJ cookies as the ones to delete). I'm wondering about whether it's excessive to expect manual cookie deletion; it's somewhat advanced for Jo-average user? I tend to think just getting them to delete cookies period will suffice - having to perhaps re-log in to some other sites is very unlikely to cause hardship (and yes I read the howto articles but I still hold that view).
Also, I don't know if logging out will effect the cookie deletion (can the LJ cookie not be easily/effectively deleted if it is in use - anyone??), though... chances are the user isn't logged in at the time if they've been sent to this FAQ! ;-) I assume this paragraph is just to get them to a) check if they're allowing cookies b) get them to get rid of any old ones that may be corrupt or pointing to an expired (>60 days) session on server. Similar thoughts as above apply to the FAQ #161 referral here. That FAQ discusses a specific error message and doesn't tell them how to delete cookies itself. If we're gonna point them to a document and tell them it contains info on foo then it needs to contain info on foo itself otherwise we should point them to the document the supposedly foo-containing document points to.
Suggest: Then, delete your browser's stored cookies and make sure its privacy settings are configured to properly allow cookies. Information on cookie deletion can be found at Information on making sure your browser accepts cookies can be found in your browser's help files.
(n.b. didn't include line about 'provided as a courtesy by support volunteers' as it's already above and URL is to same journal).

Incidentally I'm not sure if this info is still valid (c. 2002) and bears on the situation at hand?

You will also need to make sure that your computer's date, time, and timezone settings are correct. If your computer's clock is significantly ahead of or behind the time on LiveJournal's servers, you won't be able to remain logged in.

I see what this is saying, but reference to servers (be they in Seattle using timezone PDT or somewhere else) can make it seem more complex than it is. The time should be correct on their computer period.
Suggest: You also need to make sure your computer's date and time plus timezone settings are correct. If your computer's clock is significantly ahead of or behind the correct time for your timezone you won't be able to stay logged in.

If you are trying to use the "never expire" login mode, you may need to switch back and forth several times between "when browser closes" (session only) and "never" (never expire) when logging in. You can do this by changing the option on and then hitting Change Options.

Text should be '"Expiration: Never" login mode' as that it what appears on the login.bml page. However, as regards this paragraph, when it was discussed in the development community it was implied this suggestion was weird and would be unlikely to do anything much; no comments were made there to refute that. Again, if a user's been going through the suggestions step-by-step they'll have already removed previously existing cookies.
Since the whole being logged out issue crops up so much in Support, I'm guessing over time anything people suggested may have got lumped into the FAQ; some of it unusual? Perhaps this suggestion helps out *lots* of users of course - I am not in a position where I can say.

A short point about the "Expiration: Never" login mode' (what it does) and a link to the login page could be inserted near the top of this FAQ too but probably covered enough in that 'What're the options when I log in?' FAQ?
I rambled about the 'never' vs. 60 days expiration issue in another comment recently; same thoughts apply if the option ends up being referenced here.

Also, you may be able to remain logged in if you log in by posting a comment and checking the "login" box, rather than by logging in using the page.

IMO this basically says take advantage of bug(s) in livejournal code; if they can remain logged in in they log in from one part of the site then they should remain logged in if they log in in another. Try as I might I can't read it as 'useful insider-tip', only as 'sometimes if you log in via some parts of the site you're gonna be logged out.' As an aside, on the lj_dev thread sentiments querying this were expressed; no idea if this does work well or whether it is historic and once did, etc.

Using the "Bind to IP Address" option does not work if your IP address changes regularly (such as every time you make a new connection to your ISP). If you have selected this option and cannot remain logged into LiveJournal, unselect it when you are logging in; it may help.

I don't like the wording here.
Suggest: The "Bind to IP Address" option will not work if your IP address changes regularly. If you use a dial-up ISP, your IP address probably changes too frequently for this option to be useful. This option is off by default. If you have selected this option and cannot remain logged in, unselecting it may help.

Also, that original text uses the word 'regularly'. Will the "Bind to IP Address" option be ineffective if the IP address changes once after the session on database has been created and browser cookie pointing to it created? Is the word 'regularly' used because problems are likely if the session continually tries to update the stored IP address? That'd have a bearing on phrasing to use, hopefully someone can clarify the session-cookie behaviour.

Having written this entry I didn't write it with a Further Reading section as am still getting to grips with new documentation format, but guess any documentation links I cited can easily just be thrown together sequentially at the bottom of the FAQ in order mentioned in FAQ main body and the text changes tweaked to reflect that.
Hmm, other further reading items could be What are the options when I log in?.

OT: Also as an aside, in (cache clearing) can we insert 'with many browsers, pressing the F5 key will Refresh the page' or something? It's not true for all systems but widespread enough so that if someone was using a system it didnt work for, they'd probably already know that keystroke didn't work on it.

Edit: Did some slight comment management (e.g. screening), including on my own comments. Should this offend anyone whose comment was affected please feel free to email my username privately. Thank you kindly.

  • Post a new comment


    Comments allowed for members only

    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded